Victor made a valid inference and Stardusty made a fallacy of division. Neither made a fallacy of composition, therefore there are no fallacies of composition in this thread.
Everything is logical and rational, remember, so there are no fallacies. Have SP explain it to you. He'll be happy to write another 16 paragraphs to explain it.
"IT looks as if we need criteria for when composition is a fallacy and when it isn't. Any idea where you get those?" We make them up and if we agree on them then we reason together by convention.
The google god says: "Generative AI is experimental. Learn more The fallacy of composition is an informal fallacy that occurs when someone assumes that a part of something has the same properties as the whole."
The reformulated AFR typically starts with something like this: "1. No belief is rationally inferred if it can be fully explained in terms of nonrational causes."
So, on this premise, premise 1 of the AFR, the process of rationally inferred, or rationally derived, or rationally arrived at beliefs, cannot be explained by or composed of non-rational causes or processes or causation.
We can get rid of the negations <"no" belief> and <"non" rational> by restating the equivalent premise in positive terms.
1.Beliefs that are rationally inferred must be explained at least in part in terms of rational causes.
Here, "rational" is likely used in the broader sense, not merely an unthinking process that is subject to rational analysis, rather, a logically thinking process itself, AKA a mind.
Thus, the fallacy of composition becomes plain.
Premise 1 of the revised AFR claims that minds require minds, or at least, a complex higher rational process must be composed of parts some of which must themselves be rational.
Perhaps the parts are simplistically rational and their aggregate of simpler rationalities accounts for the larger, as it were, rationality of a rationally held human belief, on premise 1 of the AFR.
Because premise 1 depends upon a logical fallacy, the fallacy of composition, the AFR is logically invalid.
The fact that the AFR is logically invalid does not prove god does not exist, it merely makes the AFR an unsound argument for the existence of god.
The google god also says "Generative AI is experimental. Learn more The fallacy of division is a logical error that occurs when someone assumes that what is true for the whole must also be true for its individual parts." So, if you prefer that term "division" as opposed to "composition", fine. Both are applicable as once considers the AFR from various directions.
I'd say the inference is valid in this case because the property occupies space is transitive from parts to wholes. But it would be invalid for properties in general: has mass less than one gram, for example, is not transitive.
"occupies space is transitive from parts to wholes" Right, or at least it seems that way to us under ordinary circumstances. Under other circumstances, say with a collapsed star or black hole, whereas the parts seemed to occupy an incompressible space, in fact when a great many of those parts gather in near proximity to each other they compress each other and the total space occupied by the collection becomes less than the sum of the spaces occupied previously by the parts.
So, whether one formulates premise 1 in the negative sense, or the positive sense, or whether one considers non-rational parts adding up to non-rational parts or rational parts dividing to rational parts, the premise is logically invalid.
One can consider the fallacy of composition or the fallacy of division, it is just a matter of how one wishes to formulate the key assertion.
Is rationality transitive?
Merely assuming rationality must be transitive as part of premise 1 of the AFR is logically invalid.
" 1. Every molecule of the planet Saturn occupies space. 2. Therefore , Saturn occupies space. Fallacy of composition?" As a matter of logical argumentation, absent physical knowledge what what is meant by "molecule", "planet", "Saturn", "space" and if we consider each molecule in isolation from Saturn compared to each molecule as a part of Saturn then yes, that argument would be logically invalid.
But, we know a few things.
First, it is stated in the the OP that we are considering every molecule *of* Saturn, so that indicates that we are considering every molecule as they exist at the time of being part of Saturn, and it is stated as a fact that each such molecule does occupy space.
Further, from our general physics knowledge we know that occupation of space is a fairly simple additive property of solid material. There is no known lower bound for a unit of volume in space, nor is there a known upper bound for the volume of space.
So, implicit in the argument of the OP are several further assertions, for example, that for every molecule space occupation is additive. Further, the quantity is not specific in the assertion that X "occupies space". Any amount of space occupation satisfies the property "occupies space".
However, if we didn't know any of these things, and just abstracted the argument.
1.M of P has property O 2.Therefore P has property O
That would be logically fallacious, which does not prove that P does not have property O, it just means that premise 1 is insufficient to validly conclude 2.
"However, just re-phrasing someone else's argument and claiming that is a fallacy is itself a fallacy." Not when the logical form of the argument remains, which in my rephrasing, it does.
Here is an example https://iteadthomam.blogspot.com/2011/01/first-way-in-syllogistic-form.html Francisco J. Romero Carrasquillo, Ph.D. re-phrased Aquinas not just once, but twice, for each of his Five Ways. That is a perfectly legitimate analytical technique, when done in a manner that is accurate to the original form.
I agree that creating a strawman would be analytically illegitimate, but I did not create a strawman, I merely expressed the argument in equivalent logical terms.
1.It is only illegal by age to buy alcohol if you are not yet 21 years of age. 2.It is legal by age to buy alcohol if you are at least 21 years of age.
Those are equivalent ways to express the same legal age related alcohol purchase principle. That sort of re-phrasing is not a strawman, it is a legitimate way to explain and make more clear to some people what the principle in question is.
"What I mean by that is that is that it is a non-sequitor to attempt to find a fallacy of composition in Lewis' argument." I already did, or at least in the more clearly worded version of the first premise in a typical updated version of the AFR.
Lewis did not write an argument of any significant philosophical force or value, nor did he generally write such arguments. His writings are polemic, appeals to emotion, despair, a yearning for purpose and hope. After Anscombe showed some of the more glaring errors in his apologetics Lewis went to writing fiction that explored the human condition.
"1. No belief is rationally inferred if it can be fully explained in terms of nonrational causes." *~1.Beliefs that are rationally inferred must be explained at least in part in terms of rational causes.*
If you can point out how ~1 is some sort of strawman of 1 by all means please do so.
The reply might ne that it's non-rational causes at the atomic level, but rational inference on thepart of the whole brain, hence he inference from the atomic level the the brain level comits the fallalcy of composition.
But I maintaian that materialism requires that real causaation occurs at the base level. Wholissic causation only happens when the base level elements add up to something, as they do in the case of Saturn's rings. If you have a physically omniscient being who knows where all the paticles are, and knowswhata ring is, they cannot escape the conclusion that a ring is there. But mental causation and inference imply a diffeerent principle of causation, the power to choowe based on norms and evidence. If Charles Darwin was a material system, then you cannot say both that his theories were produced by the nonrational causes and by inference. These are two contradictory causal principles.
If a belief "can be fully explained in terms of nonrational causes", then it is simply fully explained regardless of levels of atoms or brains. So composition does not come into play.
And if all there is are nonrational causes, there are not any beliefs caused by inference, period.
BTW, to "know what a ring is" is to admit formal causation. A no-no to materialists.
Victor, "But I maintaian that materialism requires that real causaation occurs at the base level" Right. Other perceived sorts or causation are abstractions, names, shorthand for aggregates far too complex and composed of base level causal processes that are far too numerous and unquantifiable to be analyzed by a human being.
"But mental causation and inference imply a diffeerent principle of causation" No, "mental causation" is just a name for a sort of aggregate of base level causations.
" the power to choowe based on norms and evidence" Again, no different principle of causation, just a name for a sort of collection of base level causations.
" If Charles Darwin was a material system, then you cannot say both that his theories were produced by the nonrational causes and by inference." His theories were produced by nonrational causes that we merely name inference.
"These are two contradictory causal principles." Not contradictory at all. Inference IS a set of base level causations.
Causal levels. In the early 19C the steam engine was understood through the sensible properties of temperature and pressure. This was a causal understanding. Later our understanding was refined in terms of insensible atoms and molecules. This was causal too. But does it invalidate the earlier, 'higher level' understanding? I don't think so. So two levels of causal understanding in the physical can be consistent. In particular, a conception of mental causation as physical cannot be ruled out as inconsistent with a deeper causal understanding of the physical.
Base level causation. If we take atoms and molecules, or worse, quarks and leptons, to form the 'base level' then physics---which we can see as materialism with further details filled in---has it that there is no causation at this level at all, at least as traditionally understood. Not sure where that leaves the argument!
Mental causation. I'm not sure that 'causation' is the right way to think of how thought seems to follow thought. We use the phrase 'it follows that' to connect a thought or sentence to preceding thoughts or sentences. But this is a formal, normative relationship. It's not claiming that such a thought will occur to us if we rehearse its antecedents.
No one had reason to doubt the steam engine was a purely mechanical system, especially since engineers designed and built it. It is exactly the type of thing the early moderns thought they could fruitfully study. Their physics purposefully ignored animal motion and the mental and the physics we study today does the same. Consequently it seems to be a category error for physicists to claim physics will one day explain the one thing physics was designed to ignore. I suppose that a plumber could someday explain how we come to believe things in terms of water pressure and flow rates but I am dubious.
I'm not sure quantum physics claims there is no causation. Indeterminate outcomes does not mean no causation.
Lewis was discussing what causes us to believe something. One way is to see something and believe it is present. The immediate sensation is what causes us to believe in this case and doesn't involve inference. He is merely saying that we also are caused to believe things by inference "if p then q. p. therefore q". It's different to believe q by inference than by sight. Do you not think that your beliefs have causes?
Since "there is no such thing as a non-rational cause" and "His theories were produced by nonrational causes that we merely name inference."
It follows that his theories were produced by non-existent causes. Furthermore since inference just IS this set of non-existent causes, inference does not exist either.
Kevin, "No doubt "nonrational" and "non-rational" have different definitions"
"rational" has various definitions.
One dictionary definition of "rational" is "in accordance with logic". On that definition of "rational" there is no such thing as a nonrational or a non-rational cause, since all causation is in accordance with logic at base.
However, if the AFR wishes to use "rational" in the sense of "logical thinking mind" then premise 1 of the AFR is begging the question, merely asserting ad hoc as a premise the very sort of thing it intends to deduce, that a logical thinking mind is required to account for a logical thinking mind.
So, by either definition of "rational" premise 1 of the AFR fails, so the AFR is a failed argument straight away.
I'm trying to understand what VR means at 1:27 PM above. He says, If Charles Darwin was a material system, then you cannot say both that his theories were produced by the nonrational causes and by inference. These are two contradictory causal principles. The nonrational causes he has in mind are the motions of atoms and molecules, the 'base level' of physical causation. My take on this is either,
a) It begs the question by assuming that inference simply cannot arise from molecular motion, or,
b) No apparently causal phenomenon in structures above the base level counts as real causation.
I put forward the steam engine counter example in refutation of (b). My other two bolded points are merely to suggest that causation is the wrong context for discussing this. But these remarks can be ignored.
David, " So two levels of causal understanding" Right, "understanding". Just the names and abstractions we apply, not the causal process itself.
" there is no causation at this level at all, at least as traditionally understood." Then traditional understandings are wrong.
More than 100 years ago, in On the Notion of Cause, Russell clearly showed how badly mangled the notions of cause and effect were at that time. He proposed the remedy of ceasing to refer to cause at all, there being just the formula describing the mutual interactions.
I disagree with that proposed remedy. Causation is a real process. Causation occurs at the base level. Causation occurs in the present moment. All causation is mutual at base, as described by the gravitational force equation, electrostatic force equation, and other descriptions of the forces of nature.
a) It begs the question by assuming that inference simply cannot arise from molecular motion, or,
I'm pretty sure this is what he has in mind since Lewis starts with:
1. No belief is rationally inferred if it can be fully explained in terms of nonrational causes.
If molecular motion fully explains one's belief, then Lewis claims that it is not rationally inferred. He considers the "rational" cause and the "non-rational" cause mutually exclusive (hence explicitly naming one the contrary of the other). If this is true, then one cannot arise from the other.
I assume the type of naturalist that he was arguing against (like Russell) also held them to be contradictory.
David, "a) It begs the question by assuming that inference simply cannot arise from molecular motion, or,"
Right, if "rational" means "logical thinking mind" then premise 1 of the AFR is begging the question. Victor is using that sort of begging the question in the passage you cited.
Indeterminism requires effect without a cause. "No it doesn't."
Suppose we treat R as a random variable, but of a real physical process, say, direction of motion after a physical interaction, or any other physical process you might care to consider.
Suppose R takes on the particular value of 87.59 degrees in our chosen frame of reference, but it could randomly take on any value, 2.481, 176.348, or any other random value.
What is the cause of R taking on a particular value if the process is intrinsically random?
That is, not to ask what triggers R to take on some value, rather, what causes R to take on the particular value that it does, on the supposition that R is a truly intrinsically random physical process?
"He considers the "rational" cause and the "non-rational" cause mutually exclusive" That is the begging the question part.
The conclusion of the argument is that a rational mind is required to explain a rational mind.
The first premise asserts ad hoc, begging the question, that no rational process can be explained by a non-rational process. That is the conclusion slightly re-arranged in negation format.
You simply don't know what indeterminism means. I've learned by now this means there is no use continuing that particular discussion.
The first premise asserts ad hoc, begging the question, that no rational process can be explained by a non-rational process. That is the conclusion slightly re-arranged in negation format.
No. It asserts that a belief that can be fully explained by non-rational causes has been fully explained. If it has been fully explained by non-rational causes it has therefore been fully explained without recourse to rational inference. This seems uncontroversial.
Hi BM. Re your comment at 1:50 PM. You may be right. Perhaps that explains why VR insists that physical causation operates purely at the 'base level'. For then any seemingly causal behaviour at the level of higher structures, which is where the naturalist wants to locate inferential processes, can be ruled out as mere pseudo-causation, and not the genuine article.
Note how this topic---the nature of mental causation, rational inference, etc---fits into Sellars's clash of images scenario.
I agree it seems what Sellars sees as the 2 images.
I'm not sure I entirely understand Victor's statement regarding Darwin's theories. Darwin would need both his body and mind to come up with his theories so there are elements of both matter in motion and rational inference at work.
Too much "either/or" and not enough "both and" it seems to me for both Victor and Sellars.
bmiller "You simply don't know what indeterminism means. I've learned by now this means there is no use continuing that particular discussion." Translation, you are wrong, you know you are wrong, but you do not have the intellectual fortitude to admit it.
"Indeterminate" from the human perspective just means that human beings cannot determine the outcome in advance, or perhaps cannot measure the outcome precisely.
In that sense of the word "indeterminate" has nothing to do with being intrinsically random, it is merely a statement of human ignorance of the precise nature of an intrinsically determinate process.
However, to assert a fundamental indeterminism element in the progressions of material in the cosmos is to assert a fundamental intrinsic randomness at work in the real progressions of real material.
Intrinsic randomness requires an effect without a cause, so indeterminism in the fundamental sense, requires an effect without a cause.
To assert a random cause, or an indeterminate cause, or a stochastic cause in the intrinsic sense is to speak unintelligibly. Such an expression is self contradictory because there can be no cause whatsoever for an intrinsically random event.
bmiller, "This seems uncontroversial." The fact that the argument is begging the question is uncontroversial. All you did was repeat the begging the question.
"If it has been fully explained by non-rational causes it has therefore been fully explained without recourse to rational inference." Since rational inference is fully explained by non-rational causes we can employ rational inference in accounting for beliefs and fully explain both the belief and the process of using rational inference as due to non-rational causes.
All you did was repeat the begging the question assertion that rational inference cannot be accounted for with non-rational causes.
"SteveK said... The more SP talks, the more dumb she sounds." At least bmiller takes the time to construct convoluted fallacious arguments to support Victor's fallacious arguments.
David, "Perhaps that explains why VR insists that physical causation operates purely at the 'base level'. For then any seemingly causal behaviour at the level of higher structures, which is where the naturalist wants to locate inferential processes, can be ruled out as mere pseudo-causation, and not the genuine article." Right. Victor is asserting that on materialism one inevitably must be an eliminativist in order to express materialism coherently. Victor is correct about that.
Since rational inference is fully explained by non-rational causes we can employ rational inference in accounting for beliefs and fully explain both the belief and the process of using rational inference as due to non-rational causes.
Which means that all explanations are non-rational. But this one is especially so.
OP "1. Every molecule of the planet Saturn occupies space.
2. Therefore , Saturn occupies space. Fallacy of composition?" Yes.
M of P has property O Therefore, P has property O
That argument is fallacious.
Just because the argument, as presented, is fallacious does not mean the conclusion must be wrong, only that the argument, as presented, is insufficient to support the conclusion.
The OP is actually shorthand for a much longer and valid argument. First one needs to establish that space occupation is strictly additive for molecules.
Suppose we think of deductive inference as follows: given a set of antecedent sentences, we must come up with a consequent sentence in such a way that the consequent is guaranteed true if the antecedents are. That is, the inference must be truth-preserving. Suppose also that we focus purely on first-order languages and the predicate calculus. It turns out that in this narrow domain, inference is a purely syntactic business, a matter of juggling symbols around the logical terms 'all', 'some', 'and', 'if-then', and so on, without regard to the meanings of non-logical terms. Now, I'm not claiming that the human mind is some sort of 'syntactic engine'. Nevertheless, it would appear that in this very restricted domain, inference can be reduced to a finite system of rules which reflect the meanings of the logical terms alone. No more seems to be required. What, if any, are the implications of this?
It seems to mean that humans can construct rules and tools in restricted domains that can help them to solve problems. But it doesn't mean the tools are doing the same thing as the human any more than it means a pen is thinking about the words it is writing.
David "What, if any, are the implications of this?" Reduction of human language sentences to logical symbolic form can lead to wrong conclusions if the translation has errors, most especially, errors of omission.
Human language is rich, has many nuances, and even a simple sentence carries with it a large number of implications, assumptions, and postulates that the speaker likely does not even consider.
For example, Carrasquillo translated the text of the First Way to syllogistic form, and then to symbolic form. From that symbolic form he concluded that the First Way is a valid argument for a first mover. https://iteadthomam.blogspot.com/2011/01/first-way-in-syllogistic-form.html
Carrasquillo failed, however, to translate correctly, because he, like Aquinas, neglected the third case, that of circular causation. That neglect makes the First Way logically invalid, in spite of having been translated into symbolic form and thought, thereby, to be logically valid.
" Now, I'm not claiming that the human mind is some sort of 'syntactic engine'. " The human brain is many sort of engines is a rather messy serial parallel system.
Hello BM. One of the problems these results in mathematical logic can help us with is our notorious inability to keep to the narrow path of inferential validity. For us this is an ideal towards which we may aim but frequently fall away from. But would you say these investigations offer no rational insight into logical implication?
I agree that we sometimes make mistakes while trying to figure things out. We can tell its a mistake because there is a standard we can compare our logic against.
Your question regarding rational insight is pretty open-ended. There's a wide variety of answers that could apply as stated. Did you have some particular aspect in mind?
I have been dipping into CSL's Dangerous Idea, Kindle edition. Here are some comments on chapter 5, section 'Intentionality and the efficacy of mental states'.
In para 4, location 853 Victor says,
The problem is this: If the physical realm is causally closed, then it looks on the face of things as if it will go on its merry way regardless of what mental states exist, and if this is the case, then mental states simply do not matter with respect to what events are caused in the physical world.
This is the intuition that drives the first premise of the AFR. It would be the case if mental states and physical states were disjoint. But the materialist can say, and would want to say, that mental states are a subset of physical states, and then it is false.
A little later at location 882 he says,
Hasker argues that if materialism is true, even if one mental event can cause another event, the fact that such an event has a mental description will be irrelevant to the causal relationships it enters into. Materialism maintains that causal relationships are governed by physical laws, not by mental content and not by logical laws.
Again, true if physical state does not 'overlap' with mental state. But an electronic adder circuit works by physical change alone yet it conforms to the arithmetic law of addition. The summands physically cause their sum. Can't neural networks similarly conform to logical laws?
At the end of this section, location 908,
Besides, the mental, insofar as it is involved in rational inference, obeys a radically different set of laws from the laws of physics, namely, the laws of logic. The laws of logic do not result from the laws governing the physical order; in fact, they are supposed to apply not only to this world, with its physical characteristics, but to all possible worlds.
Indeed. The laws of logic apply to relations between symbolic representations, viz, sentences. Such representations must accommodate the contingencies of whatever possible world they find themselves in. The laws of logic constrain possible transformations of symbolic representations so as to preserve truth. They are categorially distinct from physical laws. This need not prevent a material transformer of representations from conforming to them. See previous remark.
If all mental states are merely a subset of physical states then mental states just are physical states. So all mental states are governed only by physical laws.
Because we can construct adding machines that follow physical laws it proves that adding is only a physical process governed by physical laws.
Although logical laws are categorically different than physical laws, we can "transform" material to follow the laws of the category it does not belong to.
If so, it seems to me that anything following the first observation is unnecessary.
Yes, as VR says, the physical realm goes its merry way. But mental states, being part of the physical realm, are not irrelevant. And higher order patterns in the physical can be perfectly real, including patterns reflecting the laws of arithmetic and logic.
You hold your beliefs in conformity to the evidence, perhaps, but you never believe because of the evidence. I amintain that science presupposes that we do form beliefs beacuse of evidence, so either materialism goes or science goes.
The laws of logic constrain possible transformations of symbolic representations so as to preserve truth. They are categorially distinct from physical laws. This need not prevent a material transformer of representations from conforming to them. See previous remark.
Cows are categorially different things than humans. If I put eyeglasses and a hat on a cow to make it look like a human would the cow then then have ceased to be in the cow category and was now in the human category? It seems it is either this or that you are maintaining that there really is no categorical difference in the first place regardless of appearances.
Because we can construct adding machines that follow physical laws it proves that adding is only a physical process governed by physical laws.
I believe you were saying this to ultimately conclude that this isn't true because adding involves mental states, which aren't physical states. If mental states were identical to physical states then all adding machines would add according to the laws of logic. But we know that isn't true because some adding machines don't function correctly despite all of them being physical adding machines. How can physical states have the quality of being in error? They can't.
"But we know that isn't true because some adding machines don't function correctly despite all of them being physical adding machines. How can physical states have the quality of being in error? They can't." So, SteveK, you suppose there must be a ghost in the machine, an erroneous ghost, to account for malfunctioning adding machines.
So all mental states are governed only by physical laws. Make this stronger and a little more precise: So all physical state changes are governed only by physical laws. It depends on what we mean by 'governed by'. If we mean 'causally determined by' then Yes. But if we mean 'described by' or 'conformant with', then No, because some 'higher level' description or law or relation may well be applicable. As in the case of the adder circuit. This is the gist of my steam engine comment back at 7:16 AM and my objection to the first premise of the AFR.
adding is only a physical process. Yes. Like intentionality, something has to be adding something to something, even if it's a column of numbers on paper. Addition is different. It's abstract.
Regarding categories, what I had in mind was that the objects of the laws of logic, which I take to be sentences, are of a radically different kind from the objects of the laws of physics, which I take to be material particles and fields. Nevertheless, I say a physical system can constrain the production of physical sentence tokens so as to conform with logical laws. By analogy with the electronic adder again.
Mental tail wagged by the physical dog? Yes, I think so. You may be a more confident thinker than me. Anecdote: In my first term reading maths we were handed a set of pretty elementary problems in group theory, one of which I couldn't solve. Rather than ask for a solution I resolved to continue to think about the problem at odd moments. Twenty-five years later, during a sleepless night, the proof came to me. I'm too embarrassed to say what the problem was, it was so easy.
Victor, I think of evidence, belief, etc, as physical because I agree there can be causal relations between these things, but the only causation I can make sense of is causation in the physical.
No, because some 'higher level' description or law or relation may well be applicable.
But wouldn't this 'higher level' description also turn out to be purely physical for a physicalist?
Addition is different. It's abstract.
I don't follow. Isn't addition is just adding things?
Regarding categories, what I had in mind was that the objects of the laws of logic, which I take to be sentences, are of a radically different kind from the objects of the laws of physics, which I take to be material particles and fields.
So objects exist that are not particles and fields?
Twenty-five years later, during a sleepless night, the proof came to me.
I am relieved. I thought I led the most boring life of earth but no longer. But I've never spent even 2 years dreaming about math problems :-)
We have been talking about what causes beliefs. But there is more than one sense of causing. It seems we mostly default to thinking in terms of an "efficient" cause such as one thing pushing another. But cause has another sense, that being in terms of an explanation. For instance one must have a mind to have a belief and so a mindful being must part of the cause or explanation of beliefs. A mindful being must be present to explain logic and so on. I wonder if sometimes people are talking past each other because of these different senses.
But wouldn't this 'higher level' description also turn out to be purely physical for a physicalist? Not necessarily. We have the example of the adder obeying laws of arithmetic. Couldn't beliefs qua physical entities obey the laws of logic?
Causation within beliefs is supposedly explanatory. To analyse such causation itself in terms of explanation sounds rather circular to me.
Our problem here is the shallowness of our access to the mental. Contrast this with the deep and rich theoretical understanding we have of the physical. How do we even learn mental language? It's not as if I can put one of my beliefs on the table, point to it, and say, 'That's a belief'.
I would have thought that a physicalist would not accept the existence of objects that obey non-physical laws or that physical objects could or would obey non-physical laws. It seems you accept these concepts as part of reality. If so, then there is no causal closure of the physical.
To analyse such causation itself in terms of explanation sounds rather circular to me.
Not to me. We have to explain what we think cause is in order to communicate our thoughts about causation.
It's not as if I can put one of my beliefs on the table, point to it, and say, 'That's a belief'.
I agree completely. That is why using the tools that worked so well for physicists working within their field won't work in this field. It's a category mistake to think they could or would.
If so, then there is no causal closure of the physical. Again, not necessarily. Obedience to physical laws and obedience to non-physical laws need not be exclusive, as demonstrated by the adder circuit. Perhaps one way of seeing this is to note that what is relevant here is not just the laws of physics. It's the laws applied to a particular arrangement of conducting, semiconducting, and insulating materials in space. This arrangement so constrains the dynamics of the system that only certain patterns of behaviour are possible. In this case patterns characterised by the abstraction of arithmetical addition. This has been my position all along. But Lewis's argument seems to rule it out from the very beginning.
Here is how Victor words the first premise of the AFR at location 520:
No belief is rationally inferred if it can be fully explained in terms of non-rational causes.
We can parody this in the following sentence about the electronic adder:
No sum is arithmetically inferred (obtained) if it can be fully explained in terms of non-arithmetic causes.
I say the parody is false. The output of the circuit is caused by physical processes, not arithmetic ones, and yet it is the arithmetic sum of the inputs.
Aside from the fact that it is humans with human intellects that design and build adders to do what they do and then while mentally assigning meanings to the inputs and outputs, the output of an adder is not a belief. The output of an adder is a voltage. A human may or may not think it is useful for adding numbers depending on whether they know what it was designed to do and whether it is functioning correctly or not.
In order to have a belief in the first place, the entity must be the kind of thing that is capable of having a belief in the first place. And adder is jut not that kind of entity.
Hi BM, I agree that adder outputs are not beliefs. But this isn't required for the parody argument to work. The intuition behind AFR premise one is that (A) rational belief formation is incompatible with physical causation. To assert their physical causation is to deny their logical causation. We generalise this a bit: (B) Nothing can obey both the laws of physics and the laws of some non-physical domain such as logic or arithmetic. Now we introduce the adder circuit and claim it is a counter-example to (B). If this is so then (B) is false. So at least one non-physical domain is compatible with the physical domain. This puts doubt on (A). The defender of the AFR needs to explain what is special about rational inference that renders it incompatible with physical causation.
At location 678 Victor says, The second stage, the subject of chapter five, attempts to show that in order to fit reasoning into our universe one must accept a dualism of fundamental explanations. That is, in addition to accepting physical explanations for physical events, we must also accept rational explanations as fundamental explanations for rational inferences. So I will concentrate my reading on Chap 5.
I don't agree with your re-wording of (A). If a belief is fully explained by non-rational causes it has been fully explained full stop, no? Other explanations may or may not be compatible with the belief but in so far as they are different from the non-rational explanation, then other explanations are wrong.
That is why I think your (B) goes too far. (A) doesn't rule out beliefs that cannot be fully explained by non-rational causes from being rationally inferred. Humans who do rational inference do so as embodied beings and so for humans some beliefs are explained partially by physical activity as well as rational inference.
David, Obedience to physical laws and obedience to non-physical laws need not be exclusive, as demonstrated by the adder circuit.
Consider groupings of apples falling from an apple tree: 2 apples fall from the tree at the same time that 3 apples fall, which together form a loose grouping of 5 apples.
Did the physical process of apples falling from trees perform an arithmetic sum? No, of course not. You are inferring this.
But there's a problem: there are other apples lying around, some a few feet away and some underneath other apple trees - why aren't those included in the arithmetic sum? Why aren't fallen leaves included in your inference, or other things? Those other things are part of the same physical process but you excluded them for what rational reason?
Suppose you say that you didn't include the other things because they are a different kind of thing. A "kind" is a universal abstraction that doesn't exist physically so your inference of a particular sum is relying on abstractions that the blind physical forces of nature compelled you to accept as accurate and true.
But suppose the blind physical forces of nature compelled you to infer something very different? You wouldn't know that your inference not true, you'd think it was. The blind physical forces are compelling you to infer and you accept it as true.
(A) rational belief formation is incompatible with physical causation. To assert their physical causation is to deny their logical causation.
I would reword this to say that rational belief becomes indistinguishable from irrational belief if blind physical causation fully explains every belief.
Suppose there are 6 randomly selected beliefs marked on a single 6-sided die. We let blind physical causation do it's thing so that a belief becomes apparent to our mind. Is the belief presented to our mind via the roll of the die a rationally inferred belief? It certainly seems rational, but that is only because another die was thrown that had different feelings marked on the sides. Would anyone trust this process to result in true beliefs?
Hi BM. Not sure I fully get you. But let me amend (B) thus: No causal process can be entirely physical and yet conform to the laws of some non-physical domain such as logic or arithmetic.
Hello Steve, and thank you for engaging with this.
Did the physical process of apples falling from trees perform an arithmetic sum? In a sense, Yes. Suppose the apples were of equal mass and fell with equal velocity. Then the impulse delivered to the Earth was five times that of a single apple's falling. But waiting for apples to fall from trees is hardly an effective computational unit that might further a creature's life chances.
Why aren't fallen leaves included in your inference, or other things? We could ask an analogous question about the adder circuit. Why am I ignoring its temperature, say? The answer is that temperature doesn't serve any purpose in the thing, a computer, say, in which the adder is incorporated. Whereas the inputs and outputs of the adder are connected to other parts enabling the adder to perform a useful function within the larger entity. Just as beliefs are connected to action in the world.
But suppose the blind physical forces of nature compelled you to infer something very different? By putting the blind physical forces to work within some contrived biological structure we can get something potentially life-enhancing. The materialist will probably say that the contriving is done by evolutionary processes. In the case of belief and inference he will claim that the structure will evolve towards one that makes only valid inferences, since invalid inferences from true beliefs are likely to be life diminishing. I'm not at all sure this is right, though. Inference seems to be a linguistic phenomenon, so it may well be learned, just like language itself, though the learning mechanism may be genetically transmitted. So we are in fact taught how to infer by engagement with parents, etc, and the world. I think this answer covers your 11:34 AM comment too.
The first premise only says that a belief that is fully explained by non-rational causes has been fully explained. That belief being fully explained needs no other explanation by inference or otherwise.
This being so, it does not necessarily imply that a physical process cannot also be in conformance with logic or arithmetic incidentally. But then the conformance would not be part of the explanation or cause (because it has already been fully explained).
The argument seeks to establish that there is more to the explanation of certain beliefs than simply non-rational causes. That is evident from your quotation from location 678.
(A) rational belief formation is incompatible with physical causation. To assert their physical causation is to deny their logical causation. We generalise this a bit: (B) Nothing can obey both the laws of physics and the laws of some non-physical domain such as logic or arithmetic.
I've been trying to think about our different perspectives.
I think you read the first premise as the assertion that rational beliefs can never have any physical causation involved, and so inferred beliefs cannot have any physical components involved(the purported physicalist view). View 1.
You have provided examples with the intent to disprove this view.
It appears to me that the first premise only asserts that if a belief has only physical causes and no rational causes it cannot be considered an inferred belief. View 2.
The difference is that view 2 allows for both rational and non-rational causes for inferred beliefs while view 1 does not.
I am arguing for view 2. There can be both physical and rational causes for certain beliefs such as those that can be logically inferred.
Hi BM, Yes, Lewis is expressing your view 2, I think. But there are some implicit assumptions: there is such a thing as rational causation; it operates on beliefs; it's wholly distinct from physical causation. It's the third assumption that begs the question against materialism. I have been trying to show by analogous examples that this assumption is not forced on us. The adder could be said to exhibit 'arithmetic causation'---changing an input causes a change in the sum output---yet it is wholly ('fully') physical. Can't the materialist say that beliefs are physical, or maybe patterns in the physical, and changes in belief have physical causes, ultimately, so that rational causation reduces to a perceived patterning in belief change?
Another analogy: before chemistry and physics made a convincing case for atomism and electromagnetism, we thought that billiard ball collisions were a canonical example of mechanical causation, which we took to be real and fundamental. Now we think mechanical forces between macroscopic objects are ultimately electromagnetic in origin. Yet thinking in terms of mechanical forces works perfectly well for us, at least for everyday living, though maybe not for theorising in physics.
it's wholly distinct from physical causation. It's the third assumption that begs the question against materialism.
This seems to be where we disagree. My view 2 does not hold that the causation of an inferred belief is wholly distinct from physical causation. Humans who have inferred beliefs do so as embodied physical beings and so for us the act of inference has a physical component. Both/and rather than either/or.
The adder could be said to exhibit 'arithmetic causation'---changing an input causes a change in the sum output---yet it is wholly ('fully') physical.
This adder is a artifact of a human intelligence, not something that naturally happens without an intelligent human designer. An intelligence which is capable of understanding and producing something like an adder and that can mentally assign non-physical meanings to physical states of the state of the machine. The adder itself does not understand anything, much less that it is doing arithmetic. If we are like that adder, then we don't understand what we are doing either. We wouldn't even understand that we don't understand because we would not be the kind of thing that understands period.
Can't the materialist say that beliefs are physical, or maybe patterns in the physical, and changes in belief have physical causes, ultimately, so that rational causation reduces to a perceived patterning in belief change?
I suppose that is what materialists claim and so the objection from the AFR comes into play. If a materialist believes that all beliefs can be fully explained by non-rational causes he has thereby committed himself to admitting that all his beliefs are non-rationally caused.
The intellect is immaterial but is part of the person. The body is material but is part of the person. So the person infers as both a function of his bodily facilities and intellectual facilities. Thinking involves both the body and the mind, doesn't it?
It is Cartesian dualism that made a radical separation within the person of the body and the mind as 2 separate substances. This in turn created the mind/body problem that in turn led to materialism on the one hand and idealism on the other.
bmiller Thinking involves both the body and the mind, doesn't it?
Yes and no. How's that for a muddled answer, lol? I will admit that my Cartesian hangover has faded a lot over time but it's not entirely gone. I say 'yes' because obviously we have a human body and we think. I say 'no' because if the intellect survives death then a body isn't necessary. I've read Dr. Feser on this subject and it's a difficult one for me to grasp.
bmiller, I found the article below, which helps further my understanding, and reinforces my "yes and no" answer - but not as it relates to this discussion.
Since we are discussing what is happening before death, not after, I would answer "yes" to your question. Yes, thinking does involve both the mind and body.
Yes, "thinking" as a normal complete person involves the apparatus required for normal thinking. So thinking as a living normal complete person does not seem possible if the person is no longer a living complete person.
My view 2 does not hold that the causation of an inferred belief is wholly distinct from physical causation. Sure, but you would say there has to be some rational causation involved, by implication non-physical. You disagree that it can be wholly physical.
This adder is a ...understands period. I agree. But is this a denial that the adder adds? Or exhibits 'arithmetic causation', as I put it. Suppose the adder were part of the guidance system of a missile? Incidentally, I wonder if a child learning to add a pair of multi digit numbers understands what it is doing.
If a materialist believes that all beliefs can be fully explained by non-rational causes he has thereby committed himself to admitting that all his beliefs are non-rationally caused. [non rationally inferred?] Only if he accepts AFR P1. Which he doesn't, because he thinks some of his beliefs are rationally inferred. AFR P1 again:
No belief is rationally inferred if it can be fully explained in terms of non-rational causes
This is intended as a substantive claim. It's not a definition of 'not rationally inferred', for example. There are three debatable terms: 'rationally inferred', 'fully explained', and 'non-rational cause'. For the AFR to have any force we have to agree the meanings of these terms and agree that the premise is true. Anscombe worried about 'fully explained'. But I ask, What does 'rationally inferred' mean? Can we agree on that?
Yes. If it is wholly physical then it cannot be partly non-physical.
But is this a denial that the adder adds?
I think we can say that it can be used by humans to simulate adding.
Only if he accepts AFR P1.
I don't think so. My statement is very simple:
If a materialist believes that all beliefs can be fully explained by non-rational causes he has thereby committed himself to admitting that all his beliefs are non-rationally caused.
Here is the syllogism:
All beliefs are fully non-rationally caused. A materialist has beliefs. All of a materialists beliefs are fully non-rationally caused.
It is beside the point how one defines 'rationally inferred' unless one defines it as 'non-rationally caused' in which case it just means that all beliefs really are fully non-rationally caused and so my syllogism is confirmed.
Sure, let's accept that ' fully explained by non-rational causes' means 'non-rationally caused'. AFR P1 says that, for beliefs, non-rationally caused implies not rationally inferred. Suppose the materialist accepts, for the sake of argument, that physically caused implies non-rationally caused. He will baulk at saying this implies not rationally inferred and so deny AFR P1. So maybe we disagree as to what 'rationally inferred' means. Hence my inquiry.
AFR P1 says that, for beliefs, non-rationally caused implies not rationally inferred.
Precisely it says (paraphrasing your formulation) "fully non-rationally caused", but OK.
I would have thought that no materialist would object to the premise that fully physical causation is fully non-rational. How could he unless he were to deny there is no difference between physical causation and rational causation? But if there is no difference, then there could be no such thing as an irrational act any more than there could be any such thing as a non-physical act. It would follow then that can be no logical fallacies including the LNC, right?
Well, first let me ask if you disagree with this definition I derived from the first entries under rational and infer. I'm not even sure Victor would agree with it.
I think the key term to focus on in "rational inference" is the word "rational" since it's a specific kind of inference.
Can fully non-rational entities and causes produce a rational response or rational inference? I don't think so.
Even if the non-rational entities and causes produce a rational response, it only appears to be rational on the surface. Why? Because rationality involves more than just the response. Rationality also involves how you get to the response. It must be done in a very specific way, a rational way, hence the very specific term "rational inference".
If the inference in question is entirely non-rational then by definition it cannot be a rational inference.
It must be done in a very specific way, a rational way, hence the very specific term "rational inference".
That specific way is generally understood to be "according to logic". It's a mental method of sorts and it's what separates a rational inference from other kinds of inferences. Whatever "according to logic" means, we know it isn't something that is also "fully explained by non-rationality". It's a contradiction.
The classical definition of man is "rational animal". It has traditionally been the attribute that separates men from animals which are separated from plants which are separated from inanimate objects.
So rational seems to be reserved for humans as far as material entities are concerned.
SteveK said: ... rationality involves more than just the response. Rationality also involves how you get to the response. It must be done in a very specific way, a rational way, hence the very specific term "rational inference".
David Brightly said: We need to know how to use the qualifier 'rational' in respect of inferences.
It seems that the issue at hand comes down to the matter of whether conscious thought, conscious thinking, is epiphenomenal or not. It seems that the term mental is being used to indicate or include non-epiphenomenal conscious thinking rather than just some types of neurological processes which occur in brains (in the case of humans). The issue once again comes down to whether under naturalism/physicalism/materialism conscious thinking is necessarily epiphenomenal. Clearly, a naturalism which asserts a nomological determinateness (apparently the 'closed' system) would appear to necessarily entail epiphenomenalism with regards to conscious thinking. Could there be a naturalism without nomological determinateness?
Hello Michael, I think you are asking a bigger question than the one at hand, and I'm not sure the answer to our present question will offer any guide to your question, or vice versa. We are currently focusing on how to demarcate rational from non-rational inference. Lewis has an example of a non-rational inference: the belief that all black dogs are dangerous resulting from the event of being bitten by a black dog as a child, or possibly thereby from the belief that some black dogs are dangerous. I assume he thinks that such an inference might well have a physical basis. He certainly seems to think that rational inference depends on some extra-physical human capacity, and that if there is no such capacity then the distinction between rational and non-rational inference collapses. For myself, I don't think we have been given a good argument to think that the physical cannot support the distinction. But first, the distinction itself!
The example of the dog being dangerous also has the counter-example of a man with reasonable grounds for believing the dog is dangerous in contrast to the man who has the belief non-rationally or irrationally. Both have the same belief, but have arrived at the belief by different means. One by logical deduction and the other from instinct or a subconscious phycological condition.
I assume that he is implying that the latter's belief can be said to be caused by non-rational means while the former's belief is said to be caused by rational inference. But if both are "really" caused by non-rational physical forces, then there really can be no distinction.
David Brightly said: I think you are asking a bigger question than the one at hand
Okay, but sometimes expanding the scope is helpful - even if only to find the limits of sensibility for whatever is the issue at hand.
David Brightly said: [Lewis] certainly seems to think that rational inference depends on some extra-physical human capacity
But does it have to be "extra-physical" if physicalism is the case but utter nomological determinateness is not? Hopefully, this sheds some light on why I think the issue is about - or includes - the matter of conscious thinking and whether or not it is epiphenomenal.
David Brightly said: if there is no such capacity then the distinction between rational and non-rational inference collapses.
bmiller said: the same belief ... arrived at the belief by different means. One by logical deduction and the other from instinct or a subconscious psychological condition.
I take bmiller's reference to "logical deduction" to indicate at least valid if not sound logic, and I am taking that deduction to indicate conscious thinking where such thinking is not yet necessarily so much a matter of habit or reflex. Of course, there is reason (such as via experience) to recognize that the practice of deduction can lead to habits or reflexes of thought which shape a psychological condition with patterns of thought that tend to be relatively more rational as distinguished from understandable. For instance, the black dog fear example is readily understandable; therefore, it could, in a sense, even be called reasonable to some extent although it fails to be rational. Given physicalism with utter nomological determinateness, the black dog fear is justified even though it is irrational, because in that context there is no basis for distinguishing between justification and explanation/description in terms of physics.
A related issue regards whether even physicalism with utterly nomological determinateness, in practice shall we say, avoids being dualistic? For instance, let's start with our very own eliminativist, StardustyPsyche. As noted here, there is an alleged distinction between that which exists and that which is real, a distinction which StardustyPsyche confirmed here. As I believe we have touched upon previously, there is also often at least some dualistic semantics when a physicalist asserting nomological determinateness distinguishes the mental from the physical. Granted that the cultural/linguistic context as it has developed makes it far easier to have discussions in terms of such a distinction, but that does not mean that utter nomological determinateness is well presented in terms which deny or at least do not presume that determinateness.
BM: ...then there really can be no distinction. But why ever not? Physical causation constrained to operate within suitable physical structures can behave in accordance with abstract patterns. One such set of abstract patterns we call 'rational deductive inference'.
David, But why ever not? Physical causation constrained to operate within suitable physical structures can behave in accordance with abstract patterns. One such set of abstract patterns we call 'rational deductive inference'.
Huh? If a non-rational substance is behaving rationally then the substance is rational. You just said it was!
Any constraint to physical causation would have to be just more of the same physical causation operating on physical structures according to the laws of physics which describe any and all behaviors according to materialism. All of which materialism considers non-rational. How can some subset of non-rational causation be designated with the qualifier 'rational' when it has been established that all causation is non-rational. That seems irrational to me.:-)
Leibniz's Mill seems relevant here. We walk into the Mill and as David suggests we observe that the mill is physically behaving according to the patterns we call 'rational deductive inference'. Do we conclude that the mill is *fully explained* by non-rational causes? No!
Gentlemen, we are trying to pin down the distinction between rational inference and non-rational inference. I don't think it's helping to introduce 'rational substance' into the discussion. Likewise 'rational causation' and 'non-rational causation'. I take Lewis in making this distinction to beg the question against the materialist. Let's agree, as suggested above at 7:30 AM, that by 'non-rational cause' Lewis just means physical cause. So his premise says, 'no belief is rationally inferred if it is physically caused'. This is strong enough to carry the AFR if it is true. But we still need to understand what 'rational inference' is and to ask whether physical causation really does rule it out.
Gentlemen, we are trying to pin down the distinction between rational inference and non-rational inference.
It seems you think this is the key to why Lewis is begging the question against the materialist. But how? Some materialists would claim it begs the question because there simply is no distinction but that doesn't seem to be your position.
At February 17, 2024 2:58 PM you seemed to claim that physical causation (aka non-rational causation) can behave like something called 'rational deductive inference'. I don't understand how this is different than the claim there is no distinction between 'non-rational causation' and 'rational deductive inference' (or just 'rational inference' in the original) although 'non-rational causation' would cover more than just 'rational inference'.
So how exactly does it beg the question from the point of view of materialists that share your understanding of materialism?
David, I was responding to your statement about distinction. I wasn't bringing up anything new. I thought the Leibniz Mill example might have made that clear. If you say the Mill is behaving rationally then you've eliminated the option of the Mill being fully explained in non-rational terms.
BM. No, but I do think that pinning down this distinction is essential to our evaluation of premise one, once we have reworded the premise to eliminate its begging the question. I am not saying the distinction between rational/irrational inference is the same as that between rational/irrational causation. I accept the former but not the latter. The materialist cannot accept a premise that begs the question against him, but if we can reword it in a way that is satisfactory to both dualists and materialists, then evaluation of the AFR can continue. So I make the suggestion that we reword the premise as 'no belief is rationally inferred if it is physically caused', which I can accept for the time being as not begging the question, and which I hope agrees with your intuitions. We can then investigate the meaning of 'rationally inferred', over which we probably will disagree!
SteveK, If you say the Mill is behaving rationally then you've eliminated the option of the Mill being fully explained in non-rational terms. Well, maybe. But I haven't said anything like 'the brain behaves rationally', for example. We are talking about rational inference. My position is that rational inference is a species of patterned behaviour, and patterned behaviour can be exhibited by physical systems. I gave the example of an adder circuit that behaves in accordance with an abstract pattern we call 'addition'.
So I make the suggestion that we reword the premise as 'no belief is rationally inferred if it is physically caused', which I can accept for the time being as not begging the question
If you can accept this formulation and not the other then it seems you've put your finger on the crux of the disagreement. If a phrase has to be changed then, at least to you, the argument will been changed. So perhaps we should discuss the difference is between 'physically caused' and 'non-rationally caused'. Why change the argument just in order to address a non-fundamental point? (As an aside, I disagree with your paraphrasing of the premise since you left out the 'fully explained' phrase.)
Are physical causes the same as rational causes then? If not, how is the changed phrasing something a materialist can accept while rejecting the unchanged phrasing?
I posted the above before reading your reply to SteveK. You explained your position clearly.
My position is that rational inference is a species of patterned behaviour, and patterned behaviour can be exhibited by physical systems. I gave the example of an adder circuit that behaves in accordance with an abstract pattern we call 'addition'.
But the 2 patterned behaviors are fundamentally different.
When we 'add' we think we are taking 2 numbers and deriving the sum (another number) according to some mental rules. An adder circuit does not think. It does not take numbers as inputs, apply some mental rules and come up with a third number.
Numbers have no extension in space, no mass, and no momentum and so cannot be affect or be affected by physical entities, forces or processes. So numbers cannot be an input to a physical system causing some physical output or even a numerical output. On the other hand our intellect can and does use numbers as inputs and outputs in the process of addition.
When we mentally assign symbols to a physical system, nothing has changed in the physical system itself. So a circuit does not suddenly start 'adding numbers' merely because we recognize we can use it to help us add. It is still reacting to voltage fluctuations according to physical laws whether we find it useful for some purpose or not.
And I thought I had made a generous offer :-) I say the original argument begs the question. At 1:50 PM you appeared to agree. This is not non-fundamental! But by all means re-insert the 'fully'. From my point of view it does no work.
David, 'no belief is rationally inferred if it is physically caused'
I'm stuck on the point I keep trying to make. It's about 2 contradicting realities that the materialist says is true - or perhaps it involves an equivocation on their part. Maybe I'm not understanding something correctly. That's very possible. Then again, maybe it's you :-)
Besides denying the AFR premise, what I know is that the materialist will claim everything in the AFR premise is material/physical.
The materialist will reword the AFR premise to something like this: "some beliefs are rationally inferred and physically caused", however in doing so they are also saying "some physical systems are not fully explained by non-rationality".
This is simple logic. If, as the materialist has said, a physical system is a rational system then it's not fully explained in non-rational terms. Now here's the contradiction / equivocation as I see it:
Because the materialist says that non-rationality isn't enough to fully explain some physical systems they cannot *also* say that a system of non-rational matter fully explains those same physical systems.
Okay, what am I missing here? Bmiller, feel free to reply if you see something. It's bugging me.
Hopefully you will find this relevant, but, if not, then maybe it will be at least a little interesting. You said: I gave the example of an adder circuit that behaves in accordance with an abstract pattern we call 'addition'.
A first problem with the adder circuit analogy, and I think it was bmiller who pointed it out, is that the analogy is parasitic on a designer with a mind that has not yet been shown to be itself nomologically determined. With addition or counting in mind (given that counting is essentially addition), I wondered whether some other example could be given, an example which, assuming physicalism, did not depend on a designer or mind which might not be nomologically determined. And I recalled one of the only two things I remember from a plant physiology class; the professor once said that a banana (which we reasonably regard as lacking a mind) would only flower and fruit after 45 leaves had been produced. Given its mindlessness, the banana might work as an example of purely physical addition; then again, for reasons to be discussed, the banana might not end up exhibiting purely physical addition.
Anyhow, back when I was taking that class, I took the professor at his word, but now I decided to check (albeit only very very briefly) into his assertion. Of course, for the sake of this discussion, I assume that the subject banana is a wild variety rather than one which has been developed via mindful human intervention. I found this article which notes: "Generally, a banana plant produces around 40 leaves before fruiting." That generally descriptor could mean that the leaf count differs between varieties, or it could mean that different individual plants of the same variety will fruit upon the production of slightly different numbers of leaves. But is it clear and obvious that counting/addition is actually occurring here? Or is it the case that "[g]enerally ... 40 leaves" are necessary before fruiting because of purely physio-chemical conditions? This is to say that there could very well be no counting/addition occurring with regards to the banana in and of itself; it could well be that the relevant parameters are most rightly explained in physio-chemical terms without reference to leaf (or even any other) counting.
Of course, as cultivators of bananas, humans with minds that can abstract (and keep in mind that those minds have yet to be established as nomologically determined) would find leaf counting a significant pragmatic practice even if/though the plant does no counting/addition. This also gets me to wondering whether physio-chemical processes ever depend upon abstraction in order to process or whether the very notion of abstraction is just some kind of short-hand used in place of physio-chemical description. But that matter is not for this discussion, of course.
In any event, and getting back to the analogy, there remains a lack of an example of how addition/counting or any other abstract thinking process is relevant to physicalism/materialism/naturalism if that conscious thinking is epiphenomenal.
Michael, Your banana plant example is similar to David's statement about apples.
David: "Did the physical process of apples falling from trees perform an arithmetic sum? In a sense, Yes. Suppose the apples were of equal mass and fell with equal velocity. Then the impulse delivered to the Earth was five times that of a single apple's falling."
My rebuttal to this is that rationality involves more than just the final response or output. Rationality also involves how you get to the response. Falling isn't a rational process performing an arithmetic sum. The same is true about the adding circuit. The designer of the circuit went through the rational process and the circuit is mimicking.
I agree. Since I was not clear enough, let me try to clarify. I do not think leaf counting is occurring with the banana; I do not think such abstract (mental) processes are effects in physicalism - at least any physicalism that insists upon nomologically determined minds/conscious thinking. I leave open the possibility of an asserted physicalism which does not insist that conscious thinking is nomologically determined and/or epiphenomenal. Whether anyone would want to call that physicalism is another matter.
the materialist will claim everything in the AFR premise is material/physical. Not this one! My view is that an essential aspect of rational inference is abstract, not concrete.
If, as the materialist has said, a physical system is a rational system then.... Again, not this one. I'm reluctant to apply 'rational' to objects or stuff (eg, as in non-rational matter). It's true that man is a rational animal but the sense of 'rational' here doesn't carry the same normative force as it does in 'rational inference'. The latter is an achievable ideal. Rational animals make mistakes. More reason to pursue a common understanding of 'rational inference'.
Rationality also involves how you get to the response Indeed! Let's apply that idea to inference.
The designer of the circuit went through the rational process and the circuit is mimicking. Isn't mimicry a kind of abstract correspondence, which is all I ask?
the analogy is parasitic on a designer with a mind that has not yet been shown to be itself nomologically determined. I'm not sure what role the demonstrated nomological determinedness of the designer's mind plays in this claim, but I'll assume that the claim implies that the analogy is parasitic on a designer with a mind. But I'm not sure that this derived claim is true. There is an ontological issue here and a genealogical one. I grant that the adder is an artifact. Nevertheless, even if we had no idea as to its origin, we can see that its behaviour conforms to the abstraction we call 'addition', especially if its structure can be reverse-engineered.
I too doubt that fruiting in banana plants involves any process that instantiates addition. Non-linear biochemical dynamics that attracts to 40-ish is much more likely an explanation.
Could you say a bit more about nomological determinateness and its role as you see it in this question?
Nevertheless, even if we had no idea as to its origin, we can see that its behaviour conforms to the abstraction we call 'addition', especially if its structure can be reverse-engineered.
I disagree. As I mentioned above, when we 'add' we 'add' numbers. Numbers are not physical things. Whatever the circuit is doing it is not doing anything with numbers.
David, My view is that an essential aspect of rational inference is abstract, not concrete.
Isn't the standard materialist view that *all things* are the result of material interactions? On that definition the abstract rational inference is only the result, or conclusion, of the material interaction. Because the abstract rational inference isn't a material interaction it's not a mental process that we would call thinking. The thinking is done at the level of material interaction.
BM, Or non-rational causes :-) Aha! Very Good! Numbers are not physical things. Whatever the circuit is doing it is not doing anything with numbers. OK. Would you accept that it is doing something with representations of numbers?
DB: My view is that an essential aspect of rational inference is abstract, not concrete. Let me expand on this a bit. Rational inference appears to be a process whereby a new belief arises from existing beliefs. We don't really know what a belief is. We think of a belief as something inside us, internal rather than external, classified as mental rather than physical. But whatever a belief really is we can express it in language, in a sentence, and convey it to others. And so a belief enters the physical external world. Note that a sentence is already an abstract thing because it can take the concrete form of spoken sounds or written squiggles. We can recognise that sounds and squiggles can represent the one sentence. Some people will go further and abstract from particular languages and writing systems to so-called propositions, but I will stop at sentences. Now, rational inference seems to me to be an 'operation' on patterns of sentences. So we have ascended another level of abstraction. 'All men are mortal' abstracts to 'All A are B'. 'Socrates is a man' abstracts to 'N is A'. Rational inference matches these two patterns together and produces the new pattern 'N is B', and then dropping down a level of abstraction from sentence patterns back to sentences we get to 'Socrates is mortal'. Note that inference is independent of the meanings of the categorematic words 'Socrates', 'Man', and 'mortal'. But not independent of the meaning of the syncategorematic word 'all'. If we replace 'all' by 'some', this inference does not proceed. Finally, the new sentence is de-expressed, as it were, to a new belief, just as if the sentence were told us by a trusted source. So that is a theory of rational inference of belief, in outline. The final thing to say for me as a materialist is that I cannot find a reason to think that any of these steps cannot be realised physically. Do enlighten me :-)
David, The final thing to say for me as a materialist is that I cannot find a reason to think that any of these steps cannot be realised physically. Do enlighten me :-)
I raised some issues in my previous comments and bmiller has too. Some quick points...
1) If these steps are happening physically then it's not fully explained by non-rational things. Not a problem is you don't ever say that sort of thing.
2) These "essential abstractions" cannot be physical. You said they weren't "concrete" which I interpreted similarly. That makes the process *not* entirely physical. Thoughts?
3) If it's entirely physical then it seems to me that logic must be a natural law of some kind. The physical process is guided, via forces, to follow a logical path. Is this your view? Why hasn't a physicist ever attempted to put this into an equation?
David More thoughts on the assumption that everything is physical:
A physical abstraction has no inherent meaning and no semantic relationship to other physical things. It can be an physical effect of a physical process, sure, but it cannot have meaning or even "point to" something beyond itself or what caused it.
The physical abstraction of "Socrates" has no meaning of Socrates any more than a shadow does. A shadow of Socrates doesn't capture anything about Socrates being a man or a cardboard cutout. Also, there's no relationship to the other different abstractions. They are all independent with no physical law that can manipulate them into forming the final abstraction "Socrates is mortal" while keeping other physical abstractions away.
Physical abstractions of Socrates, a dead man, the sky, trees, Fred and other abstract thoughts all bouncing around in your brain come together and form the final physical abstraction "Socrates is mortal"? Why not "Socrates is the sky" or "Fred and Socrates are dead"? Explain the physics that creates this very controlled and specific order?
You asked: Could you say a bit more about nomological determinateness and its role as you see it in this question?
First let me address or clear up the matter regarding something else you said: I'm not sure what role the demonstrated nomological determinedness of the designer's mind plays
I expect you grant that it is common to have the experience of conscious thinking seeming to be an active - an actual - cause rather than a mere dead-end effect, an effect which itself does not cause but which, at most, is some sort of inert product of an uninterruptible wholly physical process wherein conscious thinking and consciousness are only epiphenomenal and do not contribute to the continuation of the causal-chain physical process. Phenomenologically, conscious thinking is not a mere inevitable by-product of physics; hence, phenomenologically, a designer is one whose mental activity is not a mere by-product result; therefore, phenomenologically, the designer's mind is NOT presumed to be or experienced as being nomologically determined (even though the mind is constrained or contextual). The point here is that the adder circuit entails a designer, and, phenomenologically, that mind is presumed to effect what it does owing not to nomological necessity but, rather, owing to that mind facing a non-determinate situation in which the mind can design the circuit one way or another or not design the circuit (whether by choice or by failure). I brought up the banana issue as a way to try to come up with a situation which would better avoid the phenomenological experience of non-determinateness.
Nomological determinateness is a slightly more brief expression of nomologically necessary determinateness. As noted here, in the Varieties section, "Nomological determinism is the most common form of causal determinism. It is the notion that the past and the present dictate the future entirely and necessarily by rigid natural laws, that every occurrence results inevitably from prior events." Likewise, as noted in the Nomological determinism section of this article, "Probably the most common form of causal determinism is nomological determinism (sometimes also referred to as hard determinism, as physical determinism, or as metaphysical determinism) ... the view that all future states and events of the universe are dictated by prior states and events of the universe together with the prevailing laws of nature." These descriptions are reiterated here: "Nomological determinism ... is the notion that the past and the present dictate the future entirely and necessarily by rigid natural laws and that every occurrence inevitably results from prior events."
In effect, nomological determinism (maintaining the terminology from the previous paragraph) denies that there are actual possibilities; there are only necessities. Accordingly, there are no actual physical or metaphysical ifs.
So, when SteveK asks whether "logic must be a natural law of some kind" and whether the "physical process is guided, via forces, to follow a logical path", the answer would have to be a resounding NO inasmuch as actual possibilities which are denied by nomological determinism are necessary to logic. Without actual metaphysical or physical possibilities - the ifs - there is only description and there is no actual logical path.
In case the previous part of the posting does not satisfy the "say a bit more" request, let me refer back to an earlier posting. You had said, "[Lewis] certainly seems to think that rational inference depends on some extra-physical human capacity", and I replied, "But does it have to be 'extra-physical' if physicalism is the case but utter nomological determinateness is not? Hopefully, this sheds some light on why I think the issue is about - or includes - the matter of conscious thinking and whether or not it is epiphenomenal." Rational or logical inference does not in itself appear sufficient to dispense with physicalism if there can be a physicalism without nomologically necessary utter determinateness. Another way of considering this issue is in terms of whether physicalism can accommodate consciousness as actually non-epiphenomenal. That could be worthwhile, foundational common ground.
Thanks for the longer explanation and how your thoughts regarding materialism seem different than others.
Here are my thoughts:
OK. Would you accept that it is doing something with representations of numbers?
What the machine is 'doing' is changing physical states in response to external stimuli. The machine has no intent to represent anything. So, no I cannot attribute that activity to the machine. If anything is doing any 'representation' wrt numbers and physical states it is a human mind, not the machine.
The final thing to say for me as a materialist is that I cannot find a reason to think that any of these steps cannot be realised physically.
But all of those steps were performed mentally by a human mind aside from the physical part of writing/reading or talking/listening. Machines don't have beliefs, draw conclusions from premises, operate with or on numbers etc.
And so a belief enters the physical external world. Note that a sentence is already an abstract thing because it can take the concrete form of spoken sounds or written squiggles. We can recognise that sounds and squiggles can represent the one sentence.
But a person's belief does not enter the physical external world via sentences whether written or spoken. Squiggles and sounds may have, but not a belief. The belief has not been transformed from being an internal immaterial subject of the human intellect to being physically present in ink on paper. If it were, then a belief could be smeared, erased or burned. I suggest the sentence is a sign pointing to a belief and a sign can be either a physical object or a mental concept but it is not the thing being signified.
Is the process of process of signification itself just an instance of rational inference? If so, then explaining rational inference in terms of itself would seem circular.
Thank you for your thoughts, Gentlemen. I will make some responses and perhaps some alterations today. I am already regretting using the term 'abstraction'.
Thanks, Michael, I understand you better now. I agree that we experience ourselves as free and undetermined. I ascribe this to epistemic possibility rather than ontological. We just don't sense enough about ourselves to see the rails on which we run.
Physics tells us that the world is indeterminate at the smallest scales but at the scale at which we sense it the indeterminacies mostly average out to a determinate picture. They can, however, be amplified up to sensible scales. Geiger counters, for example. But I assume that our brains can be understood through classical physics and chemistry. (There have been claims that migratory birds sense the Earth's magnetic field through quantum phenomena but the usual objection to this is that brains are too warm and 'noisy' to amplify successfully these effects to the scale of nervous switchings.) I suggest that it is the classical determinateness that guarantees our ability to make rational inferences.
You say actual (metaphysical, ontological) possibilities are necessary to logic. But consider the static, possibility-less, world of a painting of coloured blobs. Logic still applies to the relations between the blobs. If a blob is blue then it has a green blob to its right, say. This is only description, yet it is description in logical terms.
What the machine is 'doing' is changing physical states in response to external stimuli. And we are doing more?
The machine has no intent to represent anything. Nor does a map. Yet it does represent the topography. Would the map in the satnav of a self-driving car (an auto-auto) not be representing the topography to the car? Will you say that it is just digital bits? But the bits have many layers of structure, of patterning, to which the car is sensitive---or if that is too anthropomorphic---to which the car reacts. This is the idea that we have discussed before that something is representational if it can guide an entity in its navigating the world. Literally, in this case.
So, no I cannot attribute that activity to the machine. If anything is doing any 'representation' wrt numbers and physical states it is a human mind, not the machine. OK. The 'derived intentionality' issue.
Machines don't have beliefs, draw conclusions from premises, operate with or on numbers etc. Not in the way we do, perhaps. But that doesn't rule out these phenomena being physical in us. That's the present question.
But a person's belief does not enter the physical external world via sentences whether written or spoken. You are right. This is philosophy not poetry. Perhaps I should have said a representation of a belief. But my original has a metaphorical truth. The only way minds have of copying belief from one to another is via speech or writing. We discussed this not so long ago, I think. And I would add that it's primarily through sentences that our beliefs are apparent to ourselves. Or rather our conscious selves perhaps. A belief that a black dog is dangerous, say, in conjunction with the appearance of said dog, can affect our bodies without conscious rehearsal of the belief. And then we sense a change in our bodies and become aware of the belief, perhaps in a secondary way.
Is the process of process of signification itself just an instance of rational inference? No, not in the narrow understanding I'm working with. The meanings of the categorematic terms in an inference are irrelevant, or so I claim.
What the machine is 'doing' is changing physical states in response to external stimuli. And we are doing more?
We are doing something different. We are adding numbers.
Not in the way we do, perhaps. But that doesn't rule out these phenomena being physical in us. That's the present question.
Numbers have no extension in space, no mass, and no momentum and so cannot be affect or be affected by physical entities, forces or processes. So the way we manipulate numbers cannot be physical either.
Regarding navigation software. It's basically no different from player piano rolls.. I don't believe player pianos have human intellects either just because someone loads it up and hits the start button.
So the way we manipulate numbers cannot be physical either. Turn this around: we don't manipulate numbers at all. We manipulate representations of numbers. Our problem is that we mainly know numbers through their representations. So we tend to confuse the two. Compare beliefs and their representations. In my case, guilty as charged. And we can also grasp and represent the patternednesses within the numbers. As to what they are...Discussion for another day!
piano rolls. It wasn't a question of intellect possession. It was a question of representation. There is a clear sense in which piano rolls represent or simply encode music.
Our problem is that we mainly know numbers through their representations.
How do you know that we know numbers mainly through their representation and not directly? That is wihout begging the question that only the physical exists.
Even if we can signify some physical state to represent some number it still must mean that there is some number there to represent in the first place. And in the case of a number, it would mean that a non-physical entity exists.
It wasn't a question of intellect possession.
But we've been asked to believe machines are doing the same things humans do such as 'adding'. My position is that they are not doing the same things as humans. Something has to be in possession of an intellect in order to claim something or the other is adding. My claim is that it is not the machine or software running it.
David, There is a clear sense in which piano rolls represent or simply encode music.
This is mimicry. The piano doesn't grasp the melody or the timing signature or anything related to music itself. It doesn't know that it's playing music so it cannot actually be *representing* music. I agree with bmiller, the possession of an intellect is required.
We manipulate representations of numbers. Our problem is that we mainly know numbers through their representations.
"Mainly know" or "fully know"? It's an important distinction. "Mainly" implies that there is a pathway to know numbers directly. Do I have knowledge of the external world around me directly, or do I only know the representation?
The only way a piano roll can represent music is if an intellect is involved in the process. Rain drops falling on a drum kit mimics music. An intellect makes music.
After wondering whether to bother with a response which would not be expected to effect (certainly) immediate resolution of differences, I have decided to offer the following in the belief that it is ontologically possible (HaHa) to be found eventually to be at least a little interesting.
You said, "I agree that we experience ourselves as free and undetermined. I ascribe this to epistemic possibility rather than ontological." Thinking in terms of categories such as epistemic and ontological can certainly be useful; I see them as akin to isolating or controlling for variables. However, these categories must fit back together - they must be compatible - just as it is the case that variables that have been isolated certainly fit back together (this might just be a personal preference, but I am inclined to thinking beyond categories). Retrospective analysis certainly expects to eventually eliminate the epistemic possibilities which constitute a particular ignorance in order to arrive at (let us say) ontological fact. The reasonableness of such an expectation derives from the even more reasonable expectation that there is no retrospective ontological non-determinateness. Eliminating epistemic possibilities in such a manner says nothing about - and does not depend upon - whether what is now the past, before it was past, was devoid of ontological possibilities/alternatives.
That being said, there is a sort of determinateness which is not incompatible with actual possibilities, and that determinateness would be a context consisting of alternative, conflicting possibilities. For example, the context consisting of the possibility that I do write tomorrow and the possibility that I do not write tomorrow would be a determinate context consisting of and compatible with actual alternative possibilities. Whereas retrospective epistemic possibilities can be reasonably presumed to indicate ignorance rather than non-epistemic, ontological non-determinateness, it does not follow necessarily from the ontological determinateness of the past that the future is determinate in a like manner.
You also remarked: "I suggest that it is the classical determinateness that guarantees our ability to make rational inferences." Ah, but that is very different than the notion that classical determinateness makes necessary our rational inferences. Determinateness is certainly regarded as necessary for the orderliness of the mindless and the inanimate to be observed and predicted, but that does not speak to the experience of conscious thinking as non-epiphenomenal. No one denies that rationality either depends on or presumes or expects orderliness or even determinate contexts as mentioned above; however, what is never demonstrated is that nomologically necessary utter determinateness is the case, and what is also never demonstrated is that rationality (including legitimate inference) is identical to or dependent upon nomologically necessary utter determinateness.
Finally, you also said, "... consider the static, possibility-less, world of a painting of coloured blobs. Logic still applies to the relations between the blobs. If a blob is blue then it has a green blob to its right, say. This is only description, yet it is description in logical terms." Well, I would never use logic and logical in that way. But that is not a big deal or a problem at all, because, as I believe I might have indicated some time ago, I do not believe any expression is necessary; I expect expressions to be alternatively expressible.
SteveK, Let me apologise. My 'rational inference is abstract' was a bit quick and glib. What I meant was that it is dependent on finding, matching, and manipulating patterns, and patterns within patterns. Brains are pretty good at this. But when we find patterns we are paying attention to some aspects and ignoring others, so we are abstracting. But the result of the abstraction has to be another concrete representational thing (sorry, BM) that can be further worked on. And so it goes on. The justification for saying this is all possible physically is (a) we can express rules of inference as pattern transformation, and (b) there are computer programs that implement these transformations in the course of finding and proving theorems.
Your point that logic must be a natural law of some kind is interesting. Given a language with terms for individuals and concepts they can fall under I would say that logic encapsulates the meanings of 'some', 'all', 'and', 'or', 'if-then', etc, which seem to arise naturally.
A physical abstraction has no inherent meaning and no semantic relationship to other physical things. I think that's right. It just represents a pattern.
They are all independent with no physical law that can manipulate them into forming the final abstraction The patterns we are talking about represent sentences by means of symbols. There is no problem making a physical system that juggles the symbols according to certain rules to make a new pattern.
How do the appropriate thoughts or beliefs come together in a controlled or specific order? A good question. The is the so-called 'frame problem' that bedevilled old-fashioned AI. Perhaps these days massive parallelism or a neural network architecture could find the right antecedents to an inference quickly. But we humans don't always find the right logical path ourselves, and we sometimes make invalid inferences.
We can physically manipulate physical objects. Objects with extension in space, mass, velocity and such. How do we physically manipulate a 'pattern'? Does it have any of those 3 attributes?
If a pattern is the shape of an object for instance how does that get into the brain? Wouldn't it get crowded in there if we were considering a cow?
Bmiller I can see how an external object creates the effect of a brain pattern. No different in my mind than an object creating a shadow, air movement pattern or an electrical disturbance of some kind. However, all of those representative symbols are empty in the sense that they don't represent much of anything about the object itself.
None of those things can tell us if Socrates is a man because there's nothing around - no "you" - to interpret the symbols correctly. Artifacts can't be known to be artifacts. Socrates without his legs is a different representative symbol that Socrates with his legs. Is that the same thing or a different thing (essence)?
How do you know that we know numbers mainly through their representation and not directly? From our childhood learning experience. After learning to chant the number line we are introduced to the symbols for single digit numbers and then two digit numbers and then the process of adding. This is knowing them through their representations as sequences of decimal digits. We tend to fall into thinking of numbers as objects. But they might as well be thought of geometrically as places along the number line, or movements back and forth along it. I think that is the main way we come to know them. A secondary way is through a mental model or image of some sort. I think of the natural numbers as a row of fence posts stretching away to infinity. A sort of model of the Peano axioms. What would it mean to know them directly? Through the senses, perhaps? Surely not.
But we've been asked to believe machines are doing the same things humans do such as 'adding'. Ah, I think I see why we are talking past each other. I was trained in maths. We math weirdos think of any function mapping a pair of things to a third thing that's both associative and commutative as an addition. We usually demand a 'zero' thing, something that when added to a thing leaves it unchanged. It's just a pattern. Lots of kinds of things can be added, not just numbers. Obviously it takes an intellect to see if the things and the function conform to the pattern. But we don't ask how the function 'works'. It's just a given or observed relation between pairs of things and third things. It's rather superficial, really.
The piano doesn't grasp the melody or the timing signature or anything related to music itself. It doesn't know that it's playing music so it cannot actually be *representing* music Where does this idea come from? This is surely way too demanding a requirement for representation. It would deny that photos, portraits, films, recordings, sheet music, etc, were representations. Besides, my claim is that it is the piano roll, CD, or whatever that counts as the representation not the piano or CD player. The latter are more like artificial sense organs or transducers.
This is knowing them through their representations as sequences of decimal digits.
Maybe I misread or misunderstood initially that you meant that we only understood numbers as their representations. But then this statement says that we know them through their representations. If we know them then we know them. Then again your further remarks leave me in confusion what you think. Maybe we should start with this definition from Wikipedia: A number is a mathematical object used to count, measure, and label.
Is this a good definition? If so then it seems they have no physical traits. How can we know something that doesn't have any physical traits from the perspective of a materialist?
Ah, I think I see why we are talking past each other. I was trained in maths.
I have been trained in digital design and test. So your example of a digital adder means a series of gates connected together in a certain manner for a specific purpose to me. Then maybe that is where we are talking past each other, but I find it odd regarding our relative positions.
The building blocks you are using from a mathematics perspective are completely detached from physical reality. There are no functions, mappings, associative or commutative laws that physically occupy space and time in the physical world. On the other hand, I have physical voltages, transistors, conductive paths and such that I can actually point to and measure with instruments. Yet it is you arguing that your own world doesn't really exist and mine does. It seems our positions should be reversed!
David, Besides, my claim is that it is the piano roll, CD, or whatever that counts as the representation not the piano or CD player.
You may have missed where I said "The only way a piano roll can represent music is if an intellect is involved in the process". I've discussed this already so I won't repeat myself.
You seemed to be in agreement with me here:
Me: "A physical abstraction has no inherent meaning and no semantic relationship to other physical things." You: "I think that's right. It just represents a pattern."
I can see how an external object creates the effect of a brain pattern. No different in my mind than an object creating a shadow, air movement pattern or an electrical disturbance of some kind.
But I think that is skipping steps.
The object presumably possesses a 'pattern' of some sort that somehow gets into the brain. What is that physically? If the pattern is the shape common to all cows, then how does something of that size get into something the size of a brain? It may seem like a silly thing to say, but I bring it up to illustrate that no one thinks a pattern is something that has physical attributes that ends up physically in the brain.
I don't see how one can even start to talk about patterns without having first assumed that patterns are immaterial things. Immaterial things are not available for materialists I presume.
bmiller, no one thinks a pattern is something that has physical attributes that ends up physically in the brain.
I don't understand. Brain pattern tests record the waveforms and various electrical activity in the brain. Those are patterns, yes? I know you are aware of this so you must be intending to say something else and I am failing to grasp your point.
Brain pattern tests record the waveforms and various electrical activity in the brain. Those are patterns, yes?
The patterns of electrical activity from the brain are not the patterns of cow shapes. If we are physically manipulating cow shapes in our brain then the cow shape is in our brain. That was my point.
But that doesn't mean that physical things don't exhibit patterns that intellects can recognize and abstract from those physical objects. It's just that patterns per se have no physical being apart from what they are patterns of. So I agree that we manipulate patterns, but the patterns apart from the objects they are patterns of. Since they have no physical being apart from their physical object we must be manipulating immaterial objects.
bmiller, I understand. I was willing to accept that the patterns got into the brain somehow so that we could focus on what I think are bigger problems for the materialist. But, yes, I agree that looking at a cow doesn't produce a cow-shaped pattern in the brain. I've brought up the issue of meaningless patterns to David and it appears he agrees, but maybe not.
Me: "A physical abstraction has no inherent meaning and no semantic relationship to other physical things."
David: "I think that's right. It just represents a pattern."
The patterns in the brain need to be interpreted. You look at an object and now the brain somehow has a pattern. What does the pattern mean? No clue. Socrates puts on a hat and coat and the brain has new pattern. It compares it to the first pattern where no hat and coat were involved and the result is the same - no clue. The brain has no idea what these things are because all the brain is working with are the physical patterns. The brain doesn't know essential properties from accidental properties either (I use the term artifact before by mistake).
Our conversation is getting into issues related to Searle's 'Chinese Room'. I've always thought that his thought experiment, actually a similar one, could be performed as an experiment using humans instead of a machine.
1) Put humans in a room. Let them observe everything in the room to generate 'brain patterns'. 2) Give them a box of symbols. These are the patterns that represent what they observed. 3) Have them arrange the symbols according to a set of pre-established rules. 4) Ask them if they know what any of the symbols mean, which object each one refers to or what the arrangement intends to communicate. (hint: they wouldn't know anything)
If conscious humans can't do this neither can a machine.
Apologies, gents, was away all day yesterday and unable to comment.
Firstly, SteveK's objection that any representation requires an intellect to interpret it. I can see why one might say this: all the common cases of representation involve things for we humans to make sense of. Yet it seems quite reasonable to me to say that the image of a fly on a frog's retina represents the fly to the frog. Or to the frog's brain. But maybe I should use a different word altogether. The best I can come up with is 'encoding' or perhaps 'model'. That feels rather more mechanical. And it seems to work for piano rolls and CDs and digital maps for self-driving cars.
The problem we are grappling with is that I have been trying to account for inference in terms of patterns yet it would seem that a physical thing would have no way of interacting with something abstract like a pattern. So what is the next best thing? Answer in next comment.
The next best thing has to be a representation of the abstraction. Or an encoding or model. We can't interact with numbers directly, but we can interact with strings of decimal or binary or whatever base digits. The marks '2', '3' in that order are a lot more concrete than the number twenty-three, I'd say. I can just about add up a pair of two-digit decimal numbers in my head and I tend to resort to mentally picturing what I'd write down on paper. Most additions with more digits than this overload my short term memory and I am obliged to write them down on paper or use a calculator.
The marks *AB, nA → nB in that order concretely model a family of valid Barbara style inferences. The marks can be understood as denoting an inference rule. Roughly, given two sentences one of which matches *AB and the other which matches nA, validly infer the sentence nB. A sentence matches *AB if it begins 'all' or 'every', followed by a concept word A and another concept word B. In order to match nA the other sentence must be a name word n followed by the same concept word A as in the former sentence. The inferred sentence is the name word n followed by the concept word B. The important point here is that this is wholly mechanical. With some details filled in about representations or encodings or modellings of words and sentences our marks could be compiled into a program that returns inferences given sentence pairs. Another important point is that inference is thus a syntactic business, not a semantic one. It matters not what the words mean, and they mean nothing to the mechanical manipulator.
This is not to say that this is how we infer. It just says that inference could be mechanical and hence physical.
SteveK says, The brain has no idea what these things are because all the brain is working with are the physical patterns. Absolutely. But a brain that worked this way and gave intelligible responses might give the impression to an observer that it knew what it was talking about. I think our conscious selves are observers of our own responses, so I think my own brain knows what it's talking about. Consciousness, though, I think a complete mystery. My arguments suggest that a language-using robotic entity capable of modelling a limited part of its environment, interacting with it, making inferences, and so on, is possible. It would have no sense of self. It would be forever sleepwalking, as it were. Nobody would be at home.
Since a purely physical thing cannot possibly interact with an immaterial pattern, then how can you, as a purely physical thing, create a physical representation, or even any representation at all, of a purely immaterial pattern?
It also seems to me that representation entails abstraction unless the representation is a complete copy in all respects. So the purely physical you, would have to somehow manipulate an immaterial abstraction (the pattern of inference) to produce a different abstraction (the proposed physical model) in order to physically begin to build this machine. It seems to me that the 'next best thing' just repeats the original problem twice.
David What bmiller just said is a very big part of the problem, and what I am attempting to do is highlight other problems. It's not like I expect anyone to have all of this figured out. Nobody does. What I expect is for an explanation to make rational sense on it's own terms.
Absolutely. But a brain that worked this way and gave intelligible responses might give the impression to an observer that it knew what it was talking about.
You just said "Absolutely" that the brain has no idea so how can a different brain in another observer know? Each brain is functioning on the same principles. You aren't being consistent.
Yet it seems quite reasonable to me to say that the image of a fly on a frog's retina represents the fly to the frog.
This is where bmiller and I intersect on the same problem. The essence of "fly-ness" is immaterial. I discussed this previously when I talked about universals vs. particulars and when discussing Socrates shadow and Socrates with/without a hat/coat. Humanness is not captured in a photograph. The photograph doesn't inform the brain what physical aspects belong to the human (essential) and which do not (accidental).
BM, You are now asking the historical question, How did something get to be the way it is? rather than the ontological question, What is it? How does the representation of the inference rule (in us, not robots built by us) come about? I think a materialist has to answer, Through some sort of Darwinian trial and error process. The lower storeys might be built by genetic variation and natural selection. The upper storeys by the process of language acquisition, which is also a kind of trial and error. We should find it significant that natural language has deductive inference built into it in such a simple way, reason and language tightly bound together, as it were. I can't say anything beyond this. What I have argued so far exposes a gap in the AFR, in my view. In effect I am saying, Why can't things be the way I describe? and the onus falls on the defender of the AFR to explain why, in order that his explanation remain the only candidate. And he too has an historical question to answer, for that matter.
SteveK, Let me step back a bit. You said that the patterns in the brain need to be interpreted. I'm not sure that's the right word. It suggests 'made sense of' in some way, and I don't think that is required (or possible). A rule of inference pattern representation rather has to acted on, executed in some way, and this is not beyond the capabilities of a machine. I'm also not sure that considerations in respect of consciousness are relevant here. I think a rational, language using machine with a kind of 'blind spot' where its self or I should be is possible. It would have no way of reflecting on its own mental activities, as it were. Language use and inference could be independent of consciousness.
I'm afraid I didn't follow what you said regarding essences, universals, and particulars. How do these ideas interact with your conception of representation? They don't figure in mine at all.
You are now asking the historical question, How did something get to be the way it is? rather than the ontological question, What is it?
I am asking 'what does it do?' because that will tell me what it is. If it doesn't do anything at all it is one sort of thing and if it grows teeth and chases gazelles it is another sort of thing. That is how I can tell a statue of a lion from a real lion. Likewise a purely material thing will not be able to do immaterial things.
In effect I am saying, Why can't things be the way I describe?
Because it appears you are ascribing immaterial operations to purely material things within a theoretical framework that disallows immaterial operations.
I may have misunderstood. You used the word 'create' so I took you to be asking how something comes about. Namely, how does a purely physical thing (me), incapable of interaction with an immaterial pattern, bring about a representation of that immaterial pattern within itself? Have I understood? My answer is by trial and error, aided by another purely physical thing that has already acquired a whole host of similar representations, namely a parent teaching me to speak and infer in a natural language.
It seems we agree that a purely physical being is incapable of interaction with an immaterial pattern. I have assumed that for a materialist, immaterial things do not exist since all that exists are material things. If there is nothing to represent, it seems there can be no representation of it. Maybe not so for you.
If the materialist allows that immaterial things exist but he cannot interact with them, then how can he be aware that they exist at all? I don't think it does any good to say he interacts with a 'representation' of this immaterial thing because the one making the representation would have to be aware of the immaterial thing in order to make (create) the representation. If one says he is aware that the immaterial thing exists because he has encountered a 'representation' of it we are arguing in a circle.
Additionally in order to make the representation one would have to perform the immaterial act of abstraction from the immaterial thing of the immaterial feature or features one wants to represent physically.
The same goes for the one who presumably gains understanding from interacting with this physical representation only in reverse. Why would he think it was representation of anything rather than just the physical thing itself?
Suppose I have four tokens on my desktop arranged in a square. Now, as a materialist or nominalist or whatever I don't want to say that some immaterial object, a so-called 'square', also exists. Nevertheless, the concrete world of material objects in space and time allows for spatial relations between objects. They are not of necessity all in the one place. There may be no 'immaterial objects' called squares yet our tokens are in a square in reality. Language forces me to use these abstract nouns. I can get only so far with adverbs---arranged squarely, say. You will then say, If there is nothing to represent, it seems there can be no representation of it. But that cannot be quite right. There is no square material object present, and therefore no square material object to be represented. Yet the squareness of the tokens' spatial arrangement is physically representable, or so I say.
You ask, If the materialist allows that immaterial things exist but he cannot interact with them, then how can he be aware that they exist at all? I don't concede the antecedent. I don't say that there is a physical realm of space and time filled with material objects, and another immaterial realm where their spatial relationships hang out, but to which I as a material object have no access. Rather I say that the spatial relationships are intrinsic to the space, time, material object system and that we can think them by means of physical representation within ourselves.
You say, the one making the representation would have to be aware of the immaterial thing in order to make (create) the representation. We are aware of the spatial relatedness between the tokens. This is given in vision---four distinct spots on the retina---and in proprioception---sensing that our arms and fingers have to be in different places in order to touch the tokens. The next bit is tricky. I have to say something like this. The possibility of representing squareness comes about through some trial and error process as we learn our language. We don't know what the word 'square' means until we have been shown examples and maybe the fourness and equal sidedness and right-angledness made explicit. So more basic pattern representations must be already in place, or at least developing, so that they can be connected into the emerging 'squareness' nexus. This is awfully hand-wavy but it does pay respect to the fact that language and understanding have to be learnt over a period of time by exposure to speech and examples. With this picture in place what you said starts to look backwards. I say we get the capacity to represent first, through the learning process, and only then can we be aware of the 'immaterial thing', the squareness in the arrangement of the tokens, that it can represent. I hope this makes at least some sense!
I wouldn't say an arrangement of tokens in the shape of a square requires the additional of an immaterial object either. But I would say that the shape has no physical existence apart from the objects. I think you agree.
But that cannot be quite right. There is no square material object present, and therefore no square material object to be represented. Yet the squareness of the tokens' spatial arrangement is physically representable, or so I say.
But I agree that there is a square material object present and that squareness is represented in that (those) objects. But 'squareness' detached from physical objects is not a physically existent thing. When you say that you can look at the tokens and end up drawing a square on a piece of paper I assume you have somehow separated the arrangement from the material objects or else you would have a 1 meter square arrangement of tokens in your head. Squareness is immaterial. Absent the immaterial mental capability to grasp immaterial 'squareness' all that would be physically present is an arrangement of tokens and ink on paper. 2 totally different physical things.
I agree with you that there are not 2 distinct realms.
With this picture in place what you said starts to look backwards.
But if I look at that picture it seems that you've conceded that 'squareness' exists. The fact that it can be represented abstractly by something physical is beside the point if this immaterial thing is said to exist and is knowable.
Objects arranged in a physical shape can be labeled a "square", however a larger amount of them arranged in a physical shape cannot also be labeled a "square" because it is a different physical shape. As bmiller said, squareness is immaterial so there's no ability for a physical system to grasp or to represent it.
The physical objects are not actually arranged in a square since they don't fit the definition of a square exactly. Something will be physically off and the physical system has no way to capture that difference via physical representation. No amount of trial and error can find something physical that doesn't exist physically.
The physical objects can be observed and therefore represented from various physical positions. Which is the correct representation or are all of them correct? The physical system has no idea. If you don't know what I'm talking about, David's physical arrangement represents a "square" from one position, a "rectangle" from another, and a "parallelogram" from another. There are countless physical representations.
I'm using scare quotes because shape concepts are not physical.
Gentlemen, some things you say in your comments on my 4:53 AM piece suggest to me that you both think I am putting forward a 'picture' theory of mental representation. This could not be more wrong. I think picture theories are hopeless.
I don't want to get into the ancient nominalist versus platonist debate about the ontology of properties or abstractions. I think a good theory in this area would explain why this debate has gone on for millennia. The idea is that we can form representations which do not represent, or at any rate do not represent objects. Example. This is a picture, yes, but I'm asking you to translate the concept of a non-representing representation into another representational mode, such as the one I claim operates in our brains and which generates representations in the course of our learning how to use a word like 'square'. What a representation does, in this context, is pattern recognition associated to a word. Thus we can learn to recognise a man, a tree, a square, and associate these recognisings with 'man', 'tree', and 'square', but there need be no abstract or immaterial things 'man-ness', 'tree-ness', or 'square-ness'. In particular, there is no perfect or exact or ideal square. Every thing is more or less square, mostly less. Also, I really should add that 'pattern recognition' here does not mean some kind of matching against an internally held image. Obviously, to learn the meaning of 'square' we have to be visually and proprioceptively presented with lots of instances of things we are told are 'square', and hopefully are square, in various sizes, orientations, distances, etc, practise drawing squares, cutting them out of paper, and so on. These experiences are individually forgotten but leave behind a physical recogniser, a representation.
In the light of this, does 4:53 AM make more sense?
David I'm not proposing a picture theory of mental representation. There are no mental images where the brain forms a "picture image" of the frog or the objects spatially arranged in the form of a square. My comments are directed at a generic physical brain pattern in whatever physical form you want to propose. I don't see anything in your 3:43pm comment that would change any of my objections.
OK. It was mainly the fact that you brought up the squareness of the arrangement of tokens that I formed my response around that particular pattern. But I don't think it changes the problem if we express the representation in a word rather than drawing a picture. We still have to acquire something external to ourselves, internalize it and express that something as a representation.
What we acquire is not a man or a square or anything else we can sense. If you say it is a representation of a man or a square you have not explained how that representation comes about, what it entails, and how we interact with it.
What a representation does, in this context, is pattern recognition associated to a word.
How does a representation do pattern recognition? Are you using representation as a verb or a noun?
Thus we can learn to recognise a man, a tree, a square, and associate these recognisings with 'man', 'tree', and 'square', but there need be no abstract or immaterial things 'man-ness', 'tree-ness', or 'square-ness'.
I think you are saying that if you can program a computer to digitize images, abstract away things like actual physical sizes, reduce 3 dimensions to 2 dimensions at the correct orientation, smell, texture etc and compare enough pictures of squares, there will be a series of bits in memory that represent the common essential leftovers and one can call squareness. Right?
This just seems like the adder argument again. The machine is not a rational agent, cannot know anything, cannot judge truth or falsity and cannot form any beliefs. It may help you find patterns in a signal as a tool, but only in the same sense that using fingers while counting helps you count or using a bright light and magnifying glass while inspecting an insect. I don't understand how this somehow means that a belief rationally inferred can also be fully explained in terms of non-rational causes.
*I hope asking you to translate the concept of a non-representing representation into another representational mode, isn't key to your argument because the wording makes my head hurt.
David, Would you say that that brain is operating in way that is similar to Searles Chinese Room that I mentioned previously? If it's different, can you explain the differences? That would help me.
SteveK, I know you asked David, but here is my take:
1) Put humans in a room. Let them observe everything in the room to generate 'brain patterns'. 2) Give them a box of symbols. These are the patterns that represent what they observed. 3) Have them arrange the symbols according to a set of pre-established rules. 4) Ask them if they know what any of the symbols mean, which object each one refers to or what the arrangement intends to communicate. (hint: they wouldn't know anything)
I would say that humans do have a 'box of symbols' in real life: language. That symbolic system is normative: humans use it according to the rules that have been set down for its use.
The meaning of a word is given by an explanation of meaning, and an explanation of meaning is a rule for the use of the word explained, a standard of correct use. That being the case, I don't understand why one would think the humans in the thought experiement couldn't use those 'symbols' to describe what they have observed since they have been provided an explanation of how they are to be used.
Thanks for providing this interesting thought experiment.
I will take a shot at the Searle-ish matter but, first, a couple of preliminary points.
Physicalists deny effectiveness to the conscious experience. Physicalists assert that consciousness is epiphenomenal, inert, non-contributing. David acknowledged as much when he said "I think our conscious selves are observers of our own responses" rather than effectors of anything. Accordingly, the fact that humans exhibit consciousness is supposedly a wholly irrelevant matter if - or maybe especially if - consciousness is nothing but observation. There need be no "pre-established rules" initially with regards to the symbols. In that case, the symbols are not inherently symbols; therefore, the box contains distinguishable objects. These objects can become symbols by something like the following:
Let us say that there are a chair, a table, and a rug in the room with two people who do not communicate via their voices. One person picks up an 'x' and places it on the chair, picks up a 'y' and places it on the table, and picks up a 'z' and positions it on the rug. 'x' becomes a symbol when the (brain of the) other person incorporates the association of 'x' and chair, then likewise for the other objects and their associated symbol objects. The two people can increase the number of symbols employed in something like the following manner. The second person puts the chair on the table and places the 'x' on top of the 'y' to indicate the chair is on the table. Then one of the people gets the idea to introduce another symbol ('+') to represent 'on top of'; so, 'x' '+' 'y' symbolizes chair on top of table. Then the '-' gets used to show 'y' '-' 'x' to indicate the table is under the chair, etc., etc. At some point, it may be necessary to create a new object to serve some symbolic function, but that object only becomes effective as symbol via incorporation by the other brains using the symbols/signs/representations. Here the input to the brain is by means of light waves; the supposed pattern can be further confirmed by placing the chair on top of the rug and the brain thinks it sees agreement on the part of the other person when, after that placement, the other person organizes the symbols as 'x' '+' 'z' (likewise, 'z' '+' 'x' or 'x' '-' 'z' when the rug is placed on top of the chair).
The above should be sufficient to indicate that StardustyPsyche (SP) is correct to say that, from a physicalist perspective, "'rational inference' and 'mind' are just names [symbols] applied to sets of non-rational causes."
One interesting tangent to the above regards the matter of genius - the generating of something new, such as a new symbol or even the notion of communicating via symbols. According to physicalism, the new is inevitable, nomologically necessary. The non-reductive attributing of activities to a brain is a matter of convenience, because the actual activities are a far more reductive matter. SP is on point when he says, "Victor is asserting that on materialism one inevitably must be an eliminativist in order to express materialism coherently. Victor is correct about that." A thoroughly coherent physicalist explanation requires expressions in far more reductive terms than can be (dare I say ever will be) provided. Of course, coherence is not a problem unique to physicalism.
The ultimate problem for physicalism - which is to say eliminativism - regards what is left after all the eliminating. This problem is typically avoided by resorting to what I have described as dualistic thinking. That does not have to mean substance dualism; it could mean something in line with SP's distinction between that which exists and that which is real wherein importance would reside in what SP refers to as the real.
That's a nice illustration of how one can teach another to use a symbolic system. Of course, any explanation of meaning (or of use) such as this one can be misunderstood. When asked to give information about what is in the room the person fails to do so then we know that further instruction is necessary.
What is happening in the brain (or the mind) is really irrelevant to knowing one has learned to use such a symbolic system.
How does a representation come about and how does it do pattern recognition?
We know from physiology that brains consist of interconnected neurons. This has led to the study of neural networks simulated in digital computers, in particular their ability to 'learn' to recognise patterns. So it's not too great a stretch to think that children learn how to associate the appearance of things of various kinds with the names of the kinds by means of real neural networks in their brains. As you say, there is a 'taking in' of something external which becomes internalised in the structure and synaptic connections of the network. I don't think it unreasonable to call this 'representation'.
Representations that don't represent.
We can learn to recognise solid square objects and call them 'squares'. It's plausible that the network that recognises solid square objects will also respond in part to a square outline or four separate objects placed at the corners of a square. But we don't have a word for this. We talk perhaps about a 'square shape' acknowledging that this term does not denote a material object. We have a representation, somewhat like the impossible trident drawing, that doesn't represent a material thing. Reflecting on this we ask ourselves, What does it represent then? Squareness? Cue two millennia of debate.
...there will be a series of bits in memory that represent the common essential leftovers and one can call squareness.
I would not put it like that. We are not talking about leftovers from some process of discarding. The recogniser is something constructed in the brain by a process of exposure to many instances of squares and non-squares.
I don't understand how this somehow means that a belief rationally inferred can also be fully explained in terms of non-rational causes.
If rational inference is pattern matching and substitution over sentences, and visual recognition is an activity possible for a neural network, it's plausible that rational inference is also possible for a neural network.
Even if I agreed with your last sentence the machine would not cause the machine to believe anything because machines are not the sort of things that can believe.
But leaving that aside there are more fundamental issues. Maybe you can help me understand.
If the only things that actually exist are fundamental particles then there really are no 'things' such as machines. There would be no principled reason to distinguish between the particles that the machine consists of and the those of the table is sits on or the air that surrounds it or even you. Why should I not think that half an adder, 1 cubic meter of air and 2 pounds of table are a thing rather than the parts of 3 separate things?
It seems the materialist allows himself the substance concept on the one hand to explain how materialism can account for "things" such as "brain" activity and on the other hand denies that there are really any sorts of things we call brains at all.
It also seems to me that all the talk about representation, learning, pattern matching etc all start with the assumption there are such things as substances and we are an instance of one.
The same goes for arguing for materialism from the standpoint of the study of physics. The subject requires a mind and a substantial being to even begin with. So it seems incoherent to argue that minds and substantial beings do not exist as shown by physics. If true, then no one then is really around to make the argument.
Michael, "There need be no "pre-established rules" initially with regards to the symbols"
The rules are intended to represent physical laws, at least that's the way I understand the thought experiment. If that is the case then in my example using humans, the rules (physical laws) are creating the symbols AND they are arranging the symbols. Rational inference involves a specific ordered arrangement of facts (represented by symbols).
In my examples I didn't intend for the humans to know the relationship between the symbols and the external objects in the room. Going back to David's adding circuit, the circuit has no knowledge of numbers or addition and yet David claims it is adding. Impossible. In your example the humans know the relationship and I don't think that is possible given materialism.
Because the rules create the specific ordered arrangement of symbols the human participants aren't performing a rational inference. It's being done for them. Using your objects in the example that you gave, the rules might create the arrangement "x y - + z". Is that a rational inference involving the facts?
Much more can be said but I will stop here for now.
BM, Is anyone claiming only fundamental particles exist? Not I. Trees, houses, and men seem to me to have a much better claim on existence than quarks and electrons! I'm happy to call such things 'substances', though I prefer 'material things' or 'material objects' just to avoid the implications of 'substance' within Aristotelian theory. I am less sure about minds as things or entities. I am more inclined to say that mind is what brains do, but this is vague and contentious.
Why do we distinguish between objects and their surroundings? We know that living things are continually exchanging matter with their environment. We are like persistent eddies in the great flux. Maybe we make the distinction because our brains tell us to, as it were.
SteveK, More specifically, I say that the adder exemplifies or realises a pattern we call 'addition'. It clearly is not a mind, that is, it does not exemplify mindedness, and all that that entails.
SteveK, Searle's target with the Chinese Room argument I take to be the 'strong AI' claim that intelligent behaviour can be achieved with a stored program digital computer, or maybe that brains work 'just like' such machines. I cannot believe the latter. Such systems are extremely fragile in the face of tiny changes to their structure. Brains have evolved and further developed from babyhood and are, in contrast, very, very robust. The only way I think a SPD computer might 'do' intelligent behaviour is by simulating a physical brain. That is, by modelling the physical processes going on in a brain just as weather forecasting systems model the atmosphere.
Thanks for the clarification regarding physical objects. It helps me distinguish your particular views from other typical versions of materialism.
In this case it seems you would agree that physical entities exist as something, over and above/different from, the parts they consist of in addition to the parts they consist of. What could that physical something be?
The rules are intended to represent physical laws ... Much more can be said but I will stop here for now.
Ah, I have a better sense of your point(s). I will take up a bit of that much more with what basically amounts to stream of consciousness.
Physicalism does not actually require laws, and physicalists do not have to insist on laws. The very notion of laws of physics really can be appreciated as a semantic convenience (even if it is one which introduces some inaccuracy). There are no laws or anything actually guiding or directing what occurs. The so-called laws are simply descriptions of what occurs regularly and predictably given particular conditions. Brief tangent: Given the block space-time notion, nothing actually occurs, and the ultimate description is all just always is. Of course, there are no descriptions any more than there are patterns if there is no awareness of occurrences. (It should be apparent that this relates to the matter of: No minds; no truths.) Patterns are generated from/by awareness so that talk of patterns is talk of awareness (and its conditions/context). If rationality including its inferences is entirely a matter of pattern recognition which abstracts away/eliminates content such that content is irrelevant to the rational pattern, then rationality and its inference are merely a matter of form (a formal matter, a formality) and are empty of anything like the awareness which generated the pattern dubbed rationality. Being able to recite a text is not identical to presenting the meaning (which is more closely associated with content rather than form/pattern). Rote memorization even with repetition is not an indication of knowledge regarding meaning/content. In and of itself, reproduction of a form/pattern (including rationality to the extent that it is merely a formal matter) is not an indicator of nor dependent upon awareness. Rationality as pure form/formality requires no understanding. However, where rationality as form(ality) is expected, then an awareness is required to do the abstracting necessary to present the awareness in the form required. Rationality as form(ality) can be mimicked far more easily than can be awareness. Under a physicalism which asserts consciousness as epiphenomenal, awareness is always other than conscious thinking.
BM, I would agree? Not entirely. For artifacts like clocks and cars the parts are objects in their own right assembled into a whole. For organic things that have grown, parts are more conceptual and are not objects in their own right. They have vague boundaries and are not cleanly separable from the whole.
I understand your distinction and I agree with you in the aspect you're addressing. But the parts I was referring to where things like atoms or whatever the more fundamental material each category consists of.
OK. For the sake of discussion let's say that the parts are molecules on the understanding that there is flux at this level too, with molecules exchanging parts of themselves with others in chemical reactions. This makes the identification of molecules as persistent individuals somewhat tricky.
So do I take you as holding that there is no such thing as a base "material"? You mentioned that we are eddies in the "flux". What does this flux consist of?
How can I tell if it is in motion if I can't tell what it is? Some thing has to be moving from on place to another, no?
I'm asking because it seems we can directly or indirectly sense some objects but not this stuff? It's not that I disagree, I'm just curious on your take.
Would I be correct to say that indetectable stuff forms into things that we can detect that look and behave in various ways according to the formation of the stuff?
What if I called the indetectable stuff prime matter and proceeded like this?
When prime matter is combined with forms we end up with different material objects that look and behave in particular ways. These combinations are what we call substances.
If the previous formulation is not controversial then this second one shouldn't be either. But I suspect it is because of the words I used.
If it is irregular then how could we know it is not by chance?
Do you mean that things sometimes happen that seem to be by chance? Or something like 2 different humans may chose different actions in similar circumstances or even that the same individual may behave differently in similar circumstances at different times.
If it is the latter then we could ask the individual(s) why they did what they did. They would presumably give us reasons. Couldn't then conclude that humans regularly behave according to their reasons?
207 comments:
1 – 200 of 207 Newer› Newest»1.Saturn has rings.
2.Therefore every part Saturn is composed of must have rings.
A Saturn with rings cannot be inferred if rings can be fully explained by non-ringed causes.
IT looks as if we need criteria for when composition is a fallacy and when it isn't. Any idea where you get those?
Victor made a valid inference and Stardusty made a fallacy of division. Neither made a fallacy of composition, therefore there are no fallacies of composition in this thread.
Everything is logical and rational, remember, so there are no fallacies. Have SP explain it to you. He'll be happy to write another 16 paragraphs to explain it.
"IT looks as if we need criteria for when composition is a fallacy and when it isn't. Any idea where you get those?"
We make them up and if we agree on them then we reason together by convention.
The google god says:
"Generative AI is experimental. Learn more
The fallacy of composition is an informal fallacy that occurs when someone assumes that a part of something has the same properties as the whole."
The reformulated AFR typically starts with something like this:
"1. No belief is rationally inferred if it can be fully explained in terms of nonrational causes."
So, on this premise, premise 1 of the AFR, the process of rationally inferred, or rationally derived, or rationally arrived at beliefs, cannot be explained by or composed of non-rational causes or processes or causation.
We can get rid of the negations <"no" belief> and <"non" rational> by restating the equivalent premise in positive terms.
1.Beliefs that are rationally inferred must be explained at least in part in terms of rational causes.
Here, "rational" is likely used in the broader sense, not merely an unthinking process that is subject to rational analysis, rather, a logically thinking process itself, AKA a mind.
Thus, the fallacy of composition becomes plain.
Premise 1 of the revised AFR claims that minds require minds, or at least, a complex higher rational process must be composed of parts some of which must themselves be rational.
Perhaps the parts are simplistically rational and their aggregate of simpler rationalities accounts for the larger, as it were, rationality of a rationally held human belief, on premise 1 of the AFR.
Because premise 1 depends upon a logical fallacy, the fallacy of composition, the AFR is logically invalid.
The fact that the AFR is logically invalid does not prove god does not exist, it merely makes the AFR an unsound argument for the existence of god.
The google god also says
"Generative AI is experimental. Learn more
The fallacy of division is a logical error that occurs when someone assumes that what is true for the whole must also be true for its individual parts."
So, if you prefer that term "division" as opposed to "composition", fine. Both are applicable as once considers the AFR from various directions.
I'd say the inference is valid in this case because the property occupies space is transitive from parts to wholes. But it would be invalid for properties in general: has mass less than one gram, for example, is not transitive.
"occupies space is transitive from parts to wholes"
Right, or at least it seems that way to us under ordinary circumstances. Under other circumstances, say with a collapsed star or black hole, whereas the parts seemed to occupy an incompressible space, in fact when a great many of those parts gather in near proximity to each other they compress each other and the total space occupied by the collection becomes less than the sum of the spaces occupied previously by the parts.
So, whether one formulates premise 1 in the negative sense, or the positive sense, or whether one considers non-rational parts adding up to non-rational parts or rational parts dividing to rational parts, the premise is logically invalid.
One can consider the fallacy of composition or the fallacy of division, it is just a matter of how one wishes to formulate the key assertion.
Is rationality transitive?
Merely assuming rationality must be transitive as part of premise 1 of the AFR is logically invalid.
" 1. Every molecule of the planet Saturn occupies space. 2. Therefore , Saturn occupies space. Fallacy of composition?"
As a matter of logical argumentation, absent physical knowledge what what is meant by "molecule", "planet", "Saturn", "space" and if we consider each molecule in isolation from Saturn compared to each molecule as a part of Saturn then yes, that argument would be logically invalid.
But, we know a few things.
First, it is stated in the the OP that we are considering every molecule *of* Saturn, so that indicates that we are considering every molecule as they exist at the time of being part of Saturn, and it is stated as a fact that each such molecule does occupy space.
Further, from our general physics knowledge we know that occupation of space is a fairly simple additive property of solid material. There is no known lower bound for a unit of volume in space, nor is there a known upper bound for the volume of space.
So, implicit in the argument of the OP are several further assertions, for example, that for every molecule space occupation is additive. Further, the quantity is not specific in the assertion that X "occupies space". Any amount of space occupation satisfies the property "occupies space".
However, if we didn't know any of these things, and just abstracted the argument.
1.M of P has property O
2.Therefore P has property O
That would be logically fallacious, which does not prove that P does not have property O, it just means that premise 1 is insufficient to validly conclude 2.
But it would be invalid for properties in general: has mass less than one gram, for example, is not transitive.
But has mass is transitive. So it depends on how one phrases things. I think that is Victor's point.
However, just re-phrasing someone else's argument and claiming that is a fallacy is itself a fallacy.
What I mean by that is that is that it is a non-sequitor to attempt to find a fallacy of composition in Lewis' argument.
"However, just re-phrasing someone else's argument and claiming that is a fallacy is itself a fallacy."
Not when the logical form of the argument remains, which in my rephrasing, it does.
Here is an example
https://iteadthomam.blogspot.com/2011/01/first-way-in-syllogistic-form.html
Francisco J. Romero Carrasquillo, Ph.D. re-phrased Aquinas not just once, but twice, for each of his Five Ways. That is a perfectly legitimate analytical technique, when done in a manner that is accurate to the original form.
I agree that creating a strawman would be analytically illegitimate, but I did not create a strawman, I merely expressed the argument in equivalent logical terms.
1.It is only illegal by age to buy alcohol if you are not yet 21 years of age.
2.It is legal by age to buy alcohol if you are at least 21 years of age.
Those are equivalent ways to express the same legal age related alcohol purchase principle. That sort of re-phrasing is not a strawman, it is a legitimate way to explain and make more clear to some people what the principle in question is.
"What I mean by that is that is that it is a non-sequitor to attempt to find a fallacy of composition in Lewis' argument."
I already did, or at least in the more clearly worded version of the first premise in a typical updated version of the AFR.
Lewis did not write an argument of any significant philosophical force or value, nor did he generally write such arguments. His writings are polemic, appeals to emotion, despair, a yearning for purpose and hope. After Anscombe showed some of the more glaring errors in his apologetics Lewis went to writing fiction that explored the human condition.
"1. No belief is rationally inferred if it can be fully explained in terms of nonrational causes."
*~1.Beliefs that are rationally inferred must be explained at least in part in terms of rational causes.*
If you can point out how ~1 is some sort of strawman of 1 by all means please do so.
There is no "fallacy of composition" in Lewis' argument because there is no argument from parts to the whole.
He is presenting a dilemma.
Beliefs can be caused either by inference or by non-rational causes. If by non-rational causes then not by inference.
The reply might ne that it's non-rational causes at the atomic level, but rational inference on thepart of the whole brain, hence he inference from the atomic level the the brain level comits the fallalcy of composition.
But I maintaian that materialism requires that real causaation occurs at the base level. Wholissic causation only happens when the base level elements add up to something, as they do in the case of Saturn's rings. If you have a physically omniscient being who knows where all the paticles are, and knowswhata ring is, they cannot escape the conclusion that a ring is there. But mental causation and inference imply a diffeerent principle of causation, the power to choowe based on norms and evidence. If Charles Darwin was a material system, then you cannot say both that his theories were produced by the nonrational causes and by inference. These are two contradictory causal principles.
If a belief "can be fully explained in terms of nonrational causes", then it is simply fully explained regardless of levels of atoms or brains. So composition does not come into play.
And if all there is are nonrational causes, there are not any beliefs caused by inference, period.
BTW, to "know what a ring is" is to admit formal causation. A no-no to materialists.
Victor,
"But I maintaian that materialism requires that real causaation occurs at the base level"
Right. Other perceived sorts or causation are abstractions, names, shorthand for aggregates far too complex and composed of base level causal processes that are far too numerous and unquantifiable to be analyzed by a human being.
"But mental causation and inference imply a diffeerent principle of causation"
No, "mental causation" is just a name for a sort of aggregate of base level causations.
" the power to choowe based on norms and evidence"
Again, no different principle of causation, just a name for a sort of collection of base level causations.
" If Charles Darwin was a material system, then you cannot say both that his theories were produced by the nonrational causes and by inference."
His theories were produced by nonrational causes that we merely name inference.
"These are two contradictory causal principles."
Not contradictory at all. Inference IS a set of base level causations.
Stardusty on January 27, 2024 at 1:18, on the Rise of Anti-Semitism thread:
Therefore there is no such thing as a non-rational cause, making the AFR a non-starter
Also Stardusty in this thread:
His theories were produced by nonrational causes that we merely name inference.
No doubt "nonrational" and "non-rational" have different definitions so they can both be true at once.
Causal levels. In the early 19C the steam engine was understood through the sensible properties of temperature and pressure. This was a causal understanding. Later our understanding was refined in terms of insensible atoms and molecules. This was causal too. But does it invalidate the earlier, 'higher level' understanding? I don't think so. So two levels of causal understanding in the physical can be consistent. In particular, a conception of mental causation as physical cannot be ruled out as inconsistent with a deeper causal understanding of the physical.
Base level causation. If we take atoms and molecules, or worse, quarks and leptons, to form the 'base level' then physics---which we can see as materialism with further details filled in---has it that there is no causation at this level at all, at least as traditionally understood. Not sure where that leaves the argument!
Mental causation. I'm not sure that 'causation' is the right way to think of how thought seems to follow thought. We use the phrase 'it follows that' to connect a thought or sentence to preceding thoughts or sentences. But this is a formal, normative relationship. It's not claiming that such a thought will occur to us if we rehearse its antecedents.
Not contradictory at all. Inference IS a set of base level causations.
Now we finally have an example of the fallacy of composition.
David,
No one had reason to doubt the steam engine was a purely mechanical system, especially since engineers designed and built it. It is exactly the type of thing the early moderns thought they could fruitfully study. Their physics purposefully ignored animal motion and the mental and the physics we study today does the same. Consequently it seems to be a category error for physicists to claim physics will one day explain the one thing physics was designed to ignore. I suppose that a plumber could someday explain how we come to believe things in terms of water pressure and flow rates but I am dubious.
I'm not sure quantum physics claims there is no causation. Indeterminate outcomes does not mean no causation.
Lewis was discussing what causes us to believe something. One way is to see something and believe it is present. The immediate sensation is what causes us to believe in this case and doesn't involve inference. He is merely saying that we also are caused to believe things by inference "if p then q. p. therefore q". It's different to believe q by inference than by sight. Do you not think that your beliefs have causes?
Kevin,
Since "there is no such thing as a non-rational cause" and
"His theories were produced by nonrational causes that we merely name inference."
It follows that his theories were produced by non-existent causes. Furthermore since inference just IS this set of non-existent causes, inference does not exist either.
QED!
"Indeterminate outcomes does not mean no causation."
Yes, it does. If there is causation then the outcome is determined by the mode of causation.
Indeterminism requires effect without a cause.
"Indeterminate cause" is an incoherent term.
Kevin,
"No doubt "nonrational" and "non-rational" have different definitions"
"rational" has various definitions.
One dictionary definition of "rational" is "in accordance with logic".
On that definition of "rational" there is no such thing as a nonrational or a non-rational cause, since all causation is in accordance with logic at base.
However, if the AFR wishes to use "rational" in the sense of "logical thinking mind" then premise 1 of the AFR is begging the question, merely asserting ad hoc as a premise the very sort of thing it intends to deduce, that a logical thinking mind is required to account for a logical thinking mind.
So, by either definition of "rational" premise 1 of the AFR fails, so the AFR is a failed argument straight away.
I'm trying to understand what VR means at 1:27 PM above. He says, If Charles Darwin was a material system, then you cannot say both that his theories were produced by the nonrational causes and by inference. These are two contradictory causal principles. The nonrational causes he has in mind are the motions of atoms and molecules, the 'base level' of physical causation. My take on this is either,
a) It begs the question by assuming that inference simply cannot arise from molecular motion, or,
b) No apparently causal phenomenon in structures above the base level counts as real causation.
I put forward the steam engine counter example in refutation of (b). My other two bolded points are merely to suggest that causation is the wrong context for discussing this. But these remarks can be ignored.
David,
" So two levels of causal understanding"
Right, "understanding". Just the names and abstractions we apply, not the causal process itself.
" there is no causation at this level at all, at least as traditionally understood."
Then traditional understandings are wrong.
More than 100 years ago, in On the Notion of Cause, Russell clearly showed how badly mangled the notions of cause and effect were at that time. He proposed the remedy of ceasing to refer to cause at all, there being just the formula describing the mutual interactions.
I disagree with that proposed remedy.
Causation is a real process.
Causation occurs at the base level.
Causation occurs in the present moment.
All causation is mutual at base, as described by the gravitational force equation, electrostatic force equation, and other descriptions of the forces of nature.
a) It begs the question by assuming that inference simply cannot arise from molecular motion, or,
I'm pretty sure this is what he has in mind since Lewis starts with:
1. No belief is rationally inferred if it can be fully explained in terms of nonrational causes.
If molecular motion fully explains one's belief, then Lewis claims that it is not rationally inferred. He considers the "rational" cause and the "non-rational" cause mutually exclusive (hence explicitly naming one the contrary of the other). If this is true, then one cannot arise from the other.
I assume the type of naturalist that he was arguing against (like Russell) also held them to be contradictory.
David,
"a) It begs the question by assuming that inference simply cannot arise from molecular motion, or,"
Right, if "rational" means "logical thinking mind" then premise 1 of the AFR is begging the question. Victor is using that sort of begging the question in the passage you cited.
Yes, it does.
No it doesn't.
Indeterminism requires effect without a cause.
No it doesn't.
"Indeterminate cause" is an incoherent term.
No it isn't. In fact it is an accurate term when I use it when wondering why you waste so much time posting about things you don't understand.
Indeterminism requires effect without a cause.
"No it doesn't."
Suppose we treat R as a random variable, but of a real physical process, say, direction of motion after a physical interaction, or any other physical process you might care to consider.
Suppose R takes on the particular value of 87.59 degrees in our chosen frame of reference, but it could randomly take on any value, 2.481, 176.348, or any other random value.
What is the cause of R taking on a particular value if the process is intrinsically random?
That is, not to ask what triggers R to take on some value, rather, what causes R to take on the particular value that it does, on the supposition that R is a truly intrinsically random physical process?
"He considers the "rational" cause and the "non-rational" cause mutually exclusive"
That is the begging the question part.
The conclusion of the argument is that a rational mind is required to explain a rational mind.
The first premise asserts ad hoc, begging the question, that no rational process can be explained by a non-rational process. That is the conclusion slightly re-arranged in negation format.
Stardusty,
You simply don't know what indeterminism means. I've learned by now this means there is no use continuing that particular discussion.
The first premise asserts ad hoc, begging the question, that no rational process can be explained by a non-rational process. That is the conclusion slightly re-arranged in negation format.
No. It asserts that a belief that can be fully explained by non-rational causes has been fully explained. If it has been fully explained by non-rational causes it has therefore been fully explained without recourse to rational inference. This seems uncontroversial.
The more SP talks, the more dumb she sounds.
Hi BM. Re your comment at 1:50 PM. You may be right. Perhaps that explains why VR insists that physical causation operates purely at the 'base level'. For then any seemingly causal behaviour at the level of higher structures, which is where the naturalist wants to locate inferential processes, can be ruled out as mere pseudo-causation, and not the genuine article.
Note how this topic---the nature of mental causation, rational inference, etc---fits into Sellars's clash of images scenario.
David,
I agree it seems what Sellars sees as the 2 images.
I'm not sure I entirely understand Victor's statement regarding Darwin's theories. Darwin would need both his body and mind to come up with his theories so there are elements of both matter in motion and rational inference at work.
Too much "either/or" and not enough "both and" it seems to me for both Victor and Sellars.
bmiller
"You simply don't know what indeterminism means. I've learned by now this means there is no use continuing that particular discussion."
Translation, you are wrong, you know you are wrong, but you do not have the intellectual fortitude to admit it.
"Indeterminate" from the human perspective just means that human beings cannot determine the outcome in advance, or perhaps cannot measure the outcome precisely.
In that sense of the word "indeterminate" has nothing to do with being intrinsically random, it is merely a statement of human ignorance of the precise nature of an intrinsically determinate process.
However, to assert a fundamental indeterminism element in the progressions of material in the cosmos is to assert a fundamental intrinsic randomness at work in the real progressions of real material.
Intrinsic randomness requires an effect without a cause, so indeterminism in the fundamental sense, requires an effect without a cause.
To assert a random cause, or an indeterminate cause, or a stochastic cause in the intrinsic sense is to speak unintelligibly. Such an expression is self contradictory because there can be no cause whatsoever for an intrinsically random event.
bmiller,
"This seems uncontroversial."
The fact that the argument is begging the question is uncontroversial. All you did was repeat the begging the question.
"If it has been fully explained by non-rational causes it has therefore been fully explained without recourse to rational inference."
Since rational inference is fully explained by non-rational causes we can employ rational inference in accounting for beliefs and fully explain both the belief and the process of using rational inference as due to non-rational causes.
All you did was repeat the begging the question assertion that rational inference cannot be accounted for with non-rational causes.
"SteveK said...
The more SP talks, the more dumb she sounds."
At least bmiller takes the time to construct convoluted fallacious arguments to support Victor's fallacious arguments.
David,
"Perhaps that explains why VR insists that physical causation operates purely at the 'base level'. For then any seemingly causal behaviour at the level of higher structures, which is where the naturalist wants to locate inferential processes, can be ruled out as mere pseudo-causation, and not the genuine article."
Right. Victor is asserting that on materialism one inevitably must be an eliminativist in order to express materialism coherently. Victor is correct about that.
Translation: Stardusty can't even successfully execute simple google searches.
Since rational inference is fully explained by non-rational causes we can employ rational inference in accounting for beliefs and fully explain both the belief and the process of using rational inference as due to non-rational causes.
Which means that all explanations are non-rational. But this one is especially so.
bmiller,
"Which means that all explanations are non-rational."
At base, yes.
"rational inference" and "mind" are just names applied to sets of non-rational causes.
OP
"1. Every molecule of the planet Saturn occupies space.
2. Therefore , Saturn occupies space.
Fallacy of composition?"
Yes.
M of P has property O
Therefore, P has property O
That argument is fallacious.
Just because the argument, as presented, is fallacious does not mean the conclusion must be wrong, only that the argument, as presented, is insufficient to support the conclusion.
The OP is actually shorthand for a much longer and valid argument.
First one needs to establish that space occupation is strictly additive for molecules.
Suppose we think of deductive inference as follows: given a set of antecedent sentences, we must come up with a consequent sentence in such a way that the consequent is guaranteed true if the antecedents are. That is, the inference must be truth-preserving. Suppose also that we focus purely on first-order languages and the predicate calculus. It turns out that in this narrow domain, inference is a purely syntactic business, a matter of juggling symbols around the logical terms 'all', 'some', 'and', 'if-then', and so on, without regard to the meanings of non-logical terms. Now, I'm not claiming that the human mind is some sort of 'syntactic engine'. Nevertheless, it would appear that in this very restricted domain, inference can be reduced to a finite system of rules which reflect the meanings of the logical terms alone. No more seems to be required. What, if any, are the implications of this?
"rational inference" and "mind" are just names applied to sets of non-rational causes.
I concede that in your case it is true.
What, if any, are the implications of this?
It seems to mean that humans can construct rules and tools in restricted domains that can help them to solve problems. But it doesn't mean the tools are doing the same thing as the human any more than it means a pen is thinking about the words it is writing.
"StardustyPsyche" is a name applied to a set of non-rational causes.
David
"What, if any, are the implications of this?"
Reduction of human language sentences to logical symbolic form can lead to wrong conclusions if the translation has errors, most especially, errors of omission.
Human language is rich, has many nuances, and even a simple sentence carries with it a large number of implications, assumptions, and postulates that the speaker likely does not even consider.
For example, Carrasquillo translated the text of the First Way to syllogistic form, and then to symbolic form. From that symbolic form he concluded that the First Way is a valid argument for a first mover.
https://iteadthomam.blogspot.com/2011/01/first-way-in-syllogistic-form.html
Carrasquillo failed, however, to translate correctly, because he, like Aquinas, neglected the third case, that of circular causation. That neglect makes the First Way logically invalid, in spite of having been translated into symbolic form and thought, thereby, to be logically valid.
" Now, I'm not claiming that the human mind is some sort of 'syntactic engine'. "
The human brain is many sort of engines is a rather messy serial parallel system.
SteveK
""StardustyPsyche" is a name applied to a set of non-rational causes."
Indeed.
Hello BM. One of the problems these results in mathematical logic can help us with is our notorious inability to keep to the narrow path of inferential validity. For us this is an ideal towards which we may aim but frequently fall away from. But would you say these investigations offer no rational insight into logical implication?
Hi David,
I agree that we sometimes make mistakes while trying to figure things out. We can tell its a mistake because there is a standard we can compare our logic against.
Your question regarding rational insight is pretty open-ended. There's a wide variety of answers that could apply as stated. Did you have some particular aspect in mind?
I have been dipping into CSL's Dangerous Idea, Kindle edition. Here are some comments on chapter 5, section 'Intentionality and the efficacy of mental states'.
In para 4, location 853 Victor says,
The problem is this: If the physical realm is causally closed, then it looks on the face of things as if it will go on its merry way regardless of what mental states exist, and if this is the case, then mental states simply do not matter with respect to what events are caused in the physical world.
This is the intuition that drives the first premise of the AFR. It would be the case if mental states and physical states were disjoint. But the materialist can say, and would want to say, that mental states are a subset of physical states, and then it is false.
A little later at location 882 he says,
Hasker argues that if materialism is true, even if one mental event can cause another event, the fact that such an event has a mental description will be irrelevant to the causal relationships it enters into. Materialism maintains that causal relationships are governed by physical laws, not by mental content and not by logical laws.
Again, true if physical state does not 'overlap' with mental state. But an electronic adder circuit works by physical change alone yet it conforms to the arithmetic law of addition. The summands physically cause their sum. Can't neural networks similarly conform to logical laws?
At the end of this section, location 908,
Besides, the mental, insofar as it is involved in rational inference, obeys a radically different set of laws from the laws of physics, namely, the laws of logic. The laws of logic do not result from the laws governing the physical order; in fact, they are supposed to apply not only to this world, with its physical characteristics, but to all possible worlds.
Indeed. The laws of logic apply to relations between symbolic representations, viz, sentences. Such representations must accommodate the contingencies of whatever possible world they find themselves in. The laws of logic constrain possible transformations of symbolic representations so as to preserve truth. They are categorially distinct from physical laws. This need not prevent a material transformer of representations from conforming to them. See previous remark.
Hi David,
Can I sum up your observations as follows:
If all mental states are merely a subset of physical states then mental states just are physical states. So all mental states are governed only by physical laws.
Because we can construct adding machines that follow physical laws it proves that adding is only a physical process governed by physical laws.
Although logical laws are categorically different than physical laws, we can "transform" material to follow the laws of the category it does not belong to.
If so, it seems to me that anything following the first observation is unnecessary.
Yes, as VR says, the physical realm goes its merry way. But mental states, being part of the physical realm, are not irrelevant. And higher order patterns in the physical can be perfectly real, including patterns reflecting the laws of arithmetic and logic.
That being so, then the mental is tail being wagged by the physical dog.
It seems the other way around to me when I type responses.
You hold your beliefs in conformity to the evidence, perhaps, but you never believe because of the evidence. I amintain that science presupposes that we do form beliefs beacuse of evidence, so either materialism goes or science goes.
David,
The laws of logic constrain possible transformations of symbolic representations so as to preserve truth. They are categorially distinct from physical laws. This need not prevent a material transformer of representations from conforming to them. See previous remark.
Cows are categorially different things than humans. If I put eyeglasses and a hat on a cow to make it look like a human would the cow then then have ceased to be in the cow category and was now in the human category? It seems it is either this or that you are maintaining that there really is no categorical difference in the first place regardless of appearances.
Because we can construct adding machines that follow physical laws it proves that adding is only a physical process governed by physical laws.
I believe you were saying this to ultimately conclude that this isn't true because adding involves mental states, which aren't physical states. If mental states were identical to physical states then all adding machines would add according to the laws of logic. But we know that isn't true because some adding machines don't function correctly despite all of them being physical adding machines. How can physical states have the quality of being in error? They can't.
"But we know that isn't true because some adding machines don't function correctly despite all of them being physical adding machines. How can physical states have the quality of being in error? They can't."
So, SteveK, you suppose there must be a ghost in the machine, an erroneous ghost, to account for malfunctioning adding machines.
Victor,
"you never believe because of the evidence."
You are begging the question, again.
On materialism there is no problem with evidence influencing beliefs, so both science and materialism hold.
BM,
So all mental states are governed only by physical laws. Make this stronger and a little more precise: So all physical state changes are governed only by physical laws. It depends on what we mean by 'governed by'. If we mean 'causally determined by' then Yes. But if we mean 'described by' or 'conformant with', then No, because some 'higher level' description or law or relation may well be applicable. As in the case of the adder circuit. This is the gist of my steam engine comment back at 7:16 AM and my objection to the first premise of the AFR.
adding is only a physical process. Yes. Like intentionality, something has to be adding something to something, even if it's a column of numbers on paper. Addition is different. It's abstract.
Regarding categories, what I had in mind was that the objects of the laws of logic, which I take to be sentences, are of a radically different kind from the objects of the laws of physics, which I take to be material particles and fields. Nevertheless, I say a physical system can constrain the production of physical sentence tokens so as to conform with logical laws. By analogy with the electronic adder again.
Mental tail wagged by the physical dog? Yes, I think so. You may be a more confident thinker than me. Anecdote: In my first term reading maths we were handed a set of pretty elementary problems in group theory, one of which I couldn't solve. Rather than ask for a solution I resolved to continue to think about the problem at odd moments. Twenty-five years later, during a sleepless night, the proof came to me. I'm too embarrassed to say what the problem was, it was so easy.
Victor, I think of evidence, belief, etc, as physical because I agree there can be causal relations between these things, but the only causation I can make sense of is causation in the physical.
David,
No, because some 'higher level' description or law or relation may well be applicable.
But wouldn't this 'higher level' description also turn out to be purely physical for a physicalist?
Addition is different. It's abstract.
I don't follow. Isn't addition is just adding things?
Regarding categories, what I had in mind was that the objects of the laws of logic, which I take to be sentences, are of a radically different kind from the objects of the laws of physics, which I take to be material particles and fields.
So objects exist that are not particles and fields?
Twenty-five years later, during a sleepless night, the proof came to me.
I am relieved. I thought I led the most boring life of earth but no longer. But I've never spent even 2 years dreaming about math problems :-)
We have been talking about what causes beliefs. But there is more than one sense of causing. It seems we mostly default to thinking in terms of an "efficient" cause such as one thing pushing another. But cause has another sense, that being in terms of an explanation. For instance one must have a mind to have a belief and so a mindful being must part of the cause or explanation of beliefs. A mindful being must be present to explain logic and so on. I wonder if sometimes people are talking past each other because of these different senses.
But wouldn't this 'higher level' description also turn out to be purely physical for a physicalist? Not necessarily. We have the example of the adder obeying laws of arithmetic. Couldn't beliefs qua physical entities obey the laws of logic?
Causation within beliefs is supposedly explanatory. To analyse such causation itself in terms of explanation sounds rather circular to me.
Our problem here is the shallowness of our access to the mental. Contrast this with the deep and rich theoretical understanding we have of the physical. How do we even learn mental language? It's not as if I can put one of my beliefs on the table, point to it, and say, 'That's a belief'.
David,
I would have thought that a physicalist would not accept the existence of objects that obey non-physical laws or that physical objects could or would obey non-physical laws. It seems you accept these concepts as part of reality. If so, then there is no causal closure of the physical.
To analyse such causation itself in terms of explanation sounds rather circular to me.
Not to me. We have to explain what we think cause is in order to communicate our thoughts about causation.
It's not as if I can put one of my beliefs on the table, point to it, and say, 'That's a belief'.
I agree completely. That is why using the tools that worked so well for physicists working within their field won't work in this field. It's a category mistake to think they could or would.
If so, then there is no causal closure of the physical. Again, not necessarily. Obedience to physical laws and obedience to non-physical laws need not be exclusive, as demonstrated by the adder circuit. Perhaps one way of seeing this is to note that what is relevant here is not just the laws of physics. It's the laws applied to a particular arrangement of conducting, semiconducting, and insulating materials in space. This arrangement so constrains the dynamics of the system that only certain patterns of behaviour are possible. In this case patterns characterised by the abstraction of arithmetical addition. This has been my position all along. But Lewis's argument seems to rule it out from the very beginning.
Here is how Victor words the first premise of the AFR at location 520:
No belief is rationally inferred if it can be fully explained in terms of non-rational causes.
We can parody this in the following sentence about the electronic adder:
No sum is arithmetically inferred (obtained) if it can be fully explained in terms of non-arithmetic causes.
I say the parody is false. The output of the circuit is caused by physical processes, not arithmetic ones, and yet it is the arithmetic sum of the inputs.
David,
Aside from the fact that it is humans with human intellects that design and build adders to do what they do and then while mentally assigning meanings to the inputs and outputs, the output of an adder is not a belief. The output of an adder is a voltage. A human may or may not think it is useful for adding numbers depending on whether they know what it was designed to do and whether it is functioning correctly or not.
In order to have a belief in the first place, the entity must be the kind of thing that is capable of having a belief in the first place. And adder is jut not that kind of entity.
Sorry about the typos. Especially the last sentence.
An adder is just not that kind of entity.
Maybe put another way is that I cannot cause a digital circuit to believe it is adding. I may be able to convince a person to believe he is adding.
Hi BM, I agree that adder outputs are not beliefs. But this isn't required for the parody argument to work. The intuition behind AFR premise one is that (A) rational belief formation is incompatible with physical causation. To assert their physical causation is to deny their logical causation. We generalise this a bit: (B) Nothing can obey both the laws of physics and the laws of some non-physical domain such as logic or arithmetic. Now we introduce the adder circuit and claim it is a counter-example to (B). If this is so then (B) is false. So at least one non-physical domain is compatible with the physical domain. This puts doubt on (A). The defender of the AFR needs to explain what is special about rational inference that renders it incompatible with physical causation.
At location 678 Victor says,
The second stage, the subject of chapter five, attempts to show that in order to fit reasoning into our universe one must accept a dualism of fundamental explanations. That is, in addition to accepting physical explanations for physical events, we must also accept rational explanations as fundamental explanations for rational inferences.
So I will concentrate my reading on Chap 5.
David,
I don't agree with your re-wording of (A). If a belief is fully explained by non-rational causes it has been fully explained full stop, no? Other explanations may or may not be compatible with the belief but in so far as they are different from the non-rational explanation, then other explanations are wrong.
That is why I think your (B) goes too far. (A) doesn't rule out beliefs that cannot be fully explained by non-rational causes from being rationally inferred. Humans who do rational inference do so as embodied beings and so for humans some beliefs are explained partially by physical activity as well as rational inference.
David,
Obedience to physical laws and obedience to non-physical laws need not be exclusive, as demonstrated by the adder circuit.
Consider groupings of apples falling from an apple tree: 2 apples fall from the tree at the same time that 3 apples fall, which together form a loose grouping of 5 apples.
Did the physical process of apples falling from trees perform an arithmetic sum? No, of course not. You are inferring this.
But there's a problem: there are other apples lying around, some a few feet away and some underneath other apple trees - why aren't those included in the arithmetic sum? Why aren't fallen leaves included in your inference, or other things? Those other things are part of the same physical process but you excluded them for what rational reason?
Suppose you say that you didn't include the other things because they are a different kind of thing. A "kind" is a universal abstraction that doesn't exist physically so your inference of a particular sum is relying on abstractions that the blind physical forces of nature compelled you to accept as accurate and true.
But suppose the blind physical forces of nature compelled you to infer something very different? You wouldn't know that your inference not true, you'd think it was. The blind physical forces are compelling you to infer and you accept it as true.
(A) rational belief formation is incompatible with physical causation. To assert their physical causation is to deny their logical causation.
I would reword this to say that rational belief becomes indistinguishable from irrational belief if blind physical causation fully explains every belief.
Suppose there are 6 randomly selected beliefs marked on a single 6-sided die. We let blind physical causation do it's thing so that a belief becomes apparent to our mind. Is the belief presented to our mind via the roll of the die a rationally inferred belief? It certainly seems rational, but that is only because another die was thrown that had different feelings marked on the sides. Would anyone trust this process to result in true beliefs?
Hi BM. Not sure I fully get you. But let me amend (B) thus: No causal process can be entirely physical and yet conform to the laws of some non-physical domain such as logic or arithmetic.
Hello Steve, and thank you for engaging with this.
Did the physical process of apples falling from trees perform an arithmetic sum? In a sense, Yes. Suppose the apples were of equal mass and fell with equal velocity. Then the impulse delivered to the Earth was five times that of a single apple's falling. But waiting for apples to fall from trees is hardly an effective computational unit that might further a creature's life chances.
Why aren't fallen leaves included in your inference, or other things? We could ask an analogous question about the adder circuit. Why am I ignoring its temperature, say? The answer is that temperature doesn't serve any purpose in the thing, a computer, say, in which the adder is incorporated. Whereas the inputs and outputs of the adder are connected to other parts enabling the adder to perform a useful function within the larger entity. Just as beliefs are connected to action in the world.
But suppose the blind physical forces of nature compelled you to infer something very different? By putting the blind physical forces to work within some contrived biological structure we can get something potentially life-enhancing. The materialist will probably say that the contriving is done by evolutionary processes. In the case of belief and inference he will claim that the structure will evolve towards one that makes only valid inferences, since invalid inferences from true beliefs are likely to be life diminishing. I'm not at all sure this is right, though. Inference seems to be a linguistic phenomenon, so it may well be learned, just like language itself, though the learning mechanism may be genetically transmitted. So we are in fact taught how to infer by engagement with parents, etc, and the world. I think this answer covers your 11:34 AM comment too.
Hi David,
The first premise only says that a belief that is fully explained by non-rational causes has been fully explained. That belief being fully explained needs no other explanation by inference or otherwise.
This being so, it does not necessarily imply that a physical process cannot also be in conformance with logic or arithmetic incidentally. But then the conformance would not be part of the explanation or cause (because it has already been fully explained).
The argument seeks to establish that there is more to the explanation of certain beliefs than simply non-rational causes. That is evident from your quotation from location 678.
Hi David,
(A) rational belief formation is incompatible with physical causation. To assert their physical causation is to deny their logical causation. We generalise this a bit: (B) Nothing can obey both the laws of physics and the laws of some non-physical domain such as logic or arithmetic.
I've been trying to think about our different perspectives.
I think you read the first premise as the assertion that rational beliefs can never have any physical causation involved, and so inferred beliefs cannot have any physical components involved(the purported physicalist view). View 1.
You have provided examples with the intent to disprove this view.
It appears to me that the first premise only asserts that if a belief has only physical causes and no rational causes it cannot be considered an inferred belief. View 2.
The difference is that view 2 allows for both rational and non-rational causes for inferred beliefs while view 1 does not.
I am arguing for view 2. There can be both physical and rational causes for certain beliefs such as those that can be logically inferred.
Hi BM, Yes, Lewis is expressing your view 2, I think. But there are some implicit assumptions: there is such a thing as rational causation; it operates on beliefs; it's wholly distinct from physical causation. It's the third assumption that begs the question against materialism. I have been trying to show by analogous examples that this assumption is not forced on us. The adder could be said to exhibit 'arithmetic causation'---changing an input causes a change in the sum output---yet it is wholly ('fully') physical. Can't the materialist say that beliefs are physical, or maybe patterns in the physical, and changes in belief have physical causes, ultimately, so that rational causation reduces to a perceived patterning in belief change?
Another analogy: before chemistry and physics made a convincing case for atomism and electromagnetism, we thought that billiard ball collisions were a canonical example of mechanical causation, which we took to be real and fundamental. Now we think mechanical forces between macroscopic objects are ultimately electromagnetic in origin. Yet thinking in terms of mechanical forces works perfectly well for us, at least for everyday living, though maybe not for theorising in physics.
Hi David,
it's wholly distinct from physical causation. It's the third assumption that begs the question against materialism.
This seems to be where we disagree. My view 2 does not hold that the causation of an inferred belief is wholly distinct from physical causation. Humans who have inferred beliefs do so as embodied physical beings and so for us the act of inference has a physical component. Both/and rather than either/or.
The adder could be said to exhibit 'arithmetic causation'---changing an input causes a change in the sum output---yet it is wholly ('fully') physical.
This adder is a artifact of a human intelligence, not something that naturally happens without an intelligent human designer. An intelligence which is capable of understanding and producing something like an adder and that can mentally assign non-physical meanings to physical states of the state of the machine. The adder itself does not understand anything, much less that it is doing arithmetic. If we are like that adder, then we don't understand what we are doing either. We wouldn't even understand that we don't understand because we would not be the kind of thing that understands period.
Can't the materialist say that beliefs are physical, or maybe patterns in the physical, and changes in belief have physical causes, ultimately, so that rational causation reduces to a perceived patterning in belief change?
I suppose that is what materialists claim and so the objection from the AFR comes into play. If a materialist believes that all beliefs can be fully explained by non-rational causes he has thereby committed himself to admitting that all his beliefs are non-rationally caused.
bmiller,
My view 2 does not hold that the causation of an inferred belief is wholly distinct from physical causation
If the intellect is immaterial then doesn't that mean it is wholly distinct from the physical?
Hi SteveK,
The intellect is immaterial but is part of the person. The body is material but is part of the person. So the person infers as both a function of his bodily facilities and intellectual facilities. Thinking involves both the body and the mind, doesn't it?
It is Cartesian dualism that made a radical separation within the person of the body and the mind as 2 separate substances. This in turn created the mind/body problem that in turn led to materialism on the one hand and idealism on the other.
bmiller
Thinking involves both the body and the mind, doesn't it?
Yes and no. How's that for a muddled answer, lol? I will admit that my Cartesian hangover has faded a lot over time but it's not entirely gone. I say 'yes' because obviously we have a human body and we think. I say 'no' because if the intellect survives death then a body isn't necessary. I've read Dr. Feser on this subject and it's a difficult one for me to grasp.
bmiller,
I found the article below, which helps further my understanding, and reinforces my "yes and no" answer - but not as it relates to this discussion.
Since we are discussing what is happening before death, not after, I would answer "yes" to your question. Yes, thinking does involve both the mind and body.
https://edwardfeser.blogspot.com/2013/10/some-questions-on-soul-part-ii.html
I think the Cartesian idea is part of our background assumptions here in the US, so it's a normal default position.
But we aren't angels and we aren't animals and we aren't even angels trapped in an animal body. Well at least I'm not an angel. :-)
SteveK,
Regarding the link you provided.
Yes, "thinking" as a normal complete person involves the apparatus required for normal thinking. So thinking as a living normal complete person does not seem possible if the person is no longer a living complete person.
BM,
My view 2 does not hold that the causation of an inferred belief is wholly distinct from physical causation. Sure, but you would say there has to be some rational causation involved, by implication non-physical. You disagree that it can be wholly physical.
This adder is a ...understands period. I agree. But is this a denial that the adder adds? Or exhibits 'arithmetic causation', as I put it. Suppose the adder were part of the guidance system of a missile? Incidentally, I wonder if a child learning to add a pair of multi digit numbers understands what it is doing.
If a materialist believes that all beliefs can be fully explained by non-rational causes he has thereby committed himself to admitting that all his beliefs are non-rationally caused. [non rationally inferred?] Only if he accepts AFR P1. Which he doesn't, because he thinks some of his beliefs are rationally inferred. AFR P1 again:
No belief is rationally inferred if it can be fully explained in terms of non-rational causes
This is intended as a substantive claim. It's not a definition of 'not rationally inferred', for example. There are three debatable terms: 'rationally inferred', 'fully explained', and 'non-rational cause'. For the AFR to have any force we have to agree the meanings of these terms and agree that the premise is true. Anscombe worried about 'fully explained'. But I ask, What does 'rationally inferred' mean? Can we agree on that?
Hi David,
You disagree that it can be wholly physical.
Yes. If it is wholly physical then it cannot be partly non-physical.
But is this a denial that the adder adds?
I think we can say that it can be used by humans to simulate adding.
Only if he accepts AFR P1.
I don't think so. My statement is very simple:
If a materialist believes that all beliefs can be fully explained by non-rational causes he has thereby committed himself to admitting that all his beliefs are non-rationally caused.
Here is the syllogism:
All beliefs are fully non-rationally caused.
A materialist has beliefs.
All of a materialists beliefs are fully non-rationally caused.
It is beside the point how one defines 'rationally inferred' unless one defines it as 'non-rationally caused' in which case it just means that all beliefs really are fully non-rationally caused and so my syllogism is confirmed.
Sure, let's accept that ' fully explained by non-rational causes' means 'non-rationally caused'. AFR P1 says that, for beliefs, non-rationally caused implies not rationally inferred. Suppose the materialist accepts, for the sake of argument, that physically caused implies non-rationally caused. He will baulk at saying this implies not rationally inferred and so deny AFR P1. So maybe we disagree as to what 'rationally inferred' means. Hence my inquiry.
AFR P1 says that, for beliefs, non-rationally caused implies not rationally inferred.
Precisely it says (paraphrasing your formulation) "fully non-rationally caused", but OK.
I would have thought that no materialist would object to the premise that fully physical causation is fully non-rational. How could he unless he were to deny there is no difference between physical causation and rational causation? But if there is no difference, then there could be no such thing as an irrational act any more than there could be any such thing as a non-physical act. It would follow then that can be no logical fallacies including the LNC, right?
Yes, I buy all that. But what does 'rationally inferred' mean? What is 'rational inference'?
Let's see what Merriam-Webster says
Rational
: having reason or understanding
Infer
: to derive as a conclusion from facts or premises
So it seems 'rational inference' is having an understanding derived as a conclusion from facts or premises.
Or perhaps it could be better stated as:
Having an understanding that one is deriving a conclusion from facts or premises.
The belief resulting from this would be having the understanding that conclusion is really true.
Can you give some examples? One rational inference and one irrational inference perhaps.
Well, first let me ask if you disagree with this definition I derived from the first entries under rational and infer. I'm not even sure Victor would agree with it.
I think the key term to focus on in "rational inference" is the word "rational" since it's a specific kind of inference.
Can fully non-rational entities and causes produce a rational response or rational inference? I don't think so.
Even if the non-rational entities and causes produce a rational response, it only appears to be rational on the surface. Why? Because rationality involves more than just the response. Rationality also involves how you get to the response. It must be done in a very specific way, a rational way, hence the very specific term "rational inference".
If the inference in question is entirely non-rational then by definition it cannot be a rational inference.
I agree with SteveK. We need to know how to use the qualifier 'rational' in respect of inferences.
It must be done in a very specific way, a rational way, hence the very specific term "rational inference".
That specific way is generally understood to be "according to logic". It's a mental method of sorts and it's what separates a rational inference from other kinds of inferences. Whatever "according to logic" means, we know it isn't something that is also "fully explained by non-rationality". It's a contradiction.
The classical definition of man is "rational animal". It has traditionally been the attribute that separates men from animals which are separated from plants which are separated from inanimate objects.
So rational seems to be reserved for humans as far as material entities are concerned.
SteveK said: ... rationality involves more than just the response. Rationality also involves how you get to the response. It must be done in a very specific way, a rational way, hence the very specific term "rational inference".
David Brightly said: We need to know how to use the qualifier 'rational' in respect of inferences.
It seems that the issue at hand comes down to the matter of whether conscious thought, conscious thinking, is epiphenomenal or not. It seems that the term mental is being used to indicate or include non-epiphenomenal conscious thinking rather than just some types of neurological processes which occur in brains (in the case of humans). The issue once again comes down to whether under naturalism/physicalism/materialism conscious thinking is necessarily epiphenomenal. Clearly, a naturalism which asserts a nomological determinateness (apparently the 'closed' system) would appear to necessarily entail epiphenomenalism with regards to conscious thinking. Could there be a naturalism without nomological determinateness?
Hello Michael, I think you are asking a bigger question than the one at hand, and I'm not sure the answer to our present question will offer any guide to your question, or vice versa. We are currently focusing on how to demarcate rational from non-rational inference. Lewis has an example of a non-rational inference: the belief that all black dogs are dangerous resulting from the event of being bitten by a black dog as a child, or possibly thereby from the belief that some black dogs are dangerous. I assume he thinks that such an inference might well have a physical basis. He certainly seems to think that rational inference depends on some extra-physical human capacity, and that if there is no such capacity then the distinction between rational and non-rational inference collapses. For myself, I don't think we have been given a good argument to think that the physical cannot support the distinction. But first, the distinction itself!
David,
The example of the dog being dangerous also has the counter-example of a man with reasonable grounds for believing the dog is dangerous in contrast to the man who has the belief non-rationally or irrationally. Both have the same belief, but have arrived at the belief by different means. One by logical deduction and the other from instinct or a subconscious phycological condition.
I assume that he is implying that the latter's belief can be said to be caused by non-rational means while the former's belief is said to be caused by rational inference. But if both are "really" caused by non-rational physical forces, then there really can be no distinction.
David Brightly said: I think you are asking a bigger question than the one at hand
Okay, but sometimes expanding the scope is helpful - even if only to find the limits of sensibility for whatever is the issue at hand.
David Brightly said: [Lewis] certainly seems to think that rational inference depends on some extra-physical human capacity
But does it have to be "extra-physical" if physicalism is the case but utter nomological determinateness is not? Hopefully, this sheds some light on why I think the issue is about - or includes - the matter of conscious thinking and whether or not it is epiphenomenal.
David Brightly said: if there is no such capacity then the distinction between rational and non-rational inference collapses.
bmiller said: the same belief ... arrived at the belief by different means. One by logical deduction and the other from instinct or a subconscious psychological condition.
I take bmiller's reference to "logical deduction" to indicate at least valid if not sound logic, and I am taking that deduction to indicate conscious thinking where such thinking is not yet necessarily so much a matter of habit or reflex. Of course, there is reason (such as via experience) to recognize that the practice of deduction can lead to habits or reflexes of thought which shape a psychological condition with patterns of thought that tend to be relatively more rational as distinguished from understandable. For instance, the black dog fear example is readily understandable; therefore, it could, in a sense, even be called reasonable to some extent although it fails to be rational. Given physicalism with utter nomological determinateness, the black dog fear is justified even though it is irrational, because in that context there is no basis for distinguishing between justification and explanation/description in terms of physics.
A related issue regards whether even physicalism with utterly nomological determinateness, in practice shall we say, avoids being dualistic? For instance, let's start with our very own eliminativist, StardustyPsyche. As noted here, there is an alleged distinction between that which exists and that which is real, a distinction which StardustyPsyche confirmed here. As I believe we have touched upon previously, there is also often at least some dualistic semantics when a physicalist asserting nomological determinateness distinguishes the mental from the physical. Granted that the cultural/linguistic context as it has developed makes it far easier to have discussions in terms of such a distinction, but that does not mean that utter nomological determinateness is well presented in terms which deny or at least do not presume that determinateness.
BM: ...then there really can be no distinction. But why ever not? Physical causation constrained to operate within suitable physical structures can behave in accordance with abstract patterns. One such set of abstract patterns we call 'rational deductive inference'.
David,
But why ever not? Physical causation constrained to operate within suitable physical structures can behave in accordance with abstract patterns. One such set of abstract patterns we call 'rational deductive inference'.
Huh? If a non-rational substance is behaving rationally then the substance is rational. You just said it was!
I'm afraid I don't follow either.
Any constraint to physical causation would have to be just more of the same physical causation operating on physical structures according to the laws of physics which describe any and all behaviors according to materialism. All of which materialism considers non-rational. How can some subset of non-rational causation be designated with the qualifier 'rational' when it has been established that all causation is non-rational. That seems irrational to me.:-)
Leibniz's Mill seems relevant here. We walk into the Mill and as David suggests we observe that the mill is physically behaving according to the patterns we call 'rational deductive inference'. Do we conclude that the mill is *fully explained* by non-rational causes? No!
Gentlemen, we are trying to pin down the distinction between rational inference and non-rational inference. I don't think it's helping to introduce 'rational substance' into the discussion. Likewise 'rational causation' and 'non-rational causation'. I take Lewis in making this distinction to beg the question against the materialist. Let's agree, as suggested above at 7:30 AM, that by 'non-rational cause' Lewis just means physical cause. So his premise says, 'no belief is rationally inferred if it is physically caused'. This is strong enough to carry the AFR if it is true. But we still need to understand what 'rational inference' is and to ask whether physical causation really does rule it out.
David,
Gentlemen, we are trying to pin down the distinction between rational inference and non-rational inference.
It seems you think this is the key to why Lewis is begging the question against the materialist. But how? Some materialists would claim it begs the question because there simply is no distinction but that doesn't seem to be your position.
At February 17, 2024 2:58 PM you seemed to claim that physical causation (aka non-rational causation) can behave like something called 'rational deductive inference'. I don't understand how this is different than the claim there is no distinction between 'non-rational causation' and 'rational deductive inference' (or just 'rational inference' in the original) although 'non-rational causation' would cover more than just 'rational inference'.
So how exactly does it beg the question from the point of view of materialists that share your understanding of materialism?
David,
I was responding to your statement about distinction. I wasn't bringing up anything new. I thought the Leibniz Mill example might have made that clear. If you say the Mill is behaving rationally then you've eliminated the option of the Mill being fully explained in non-rational terms.
BM. No, but I do think that pinning down this distinction is essential to our evaluation of premise one, once we have reworded the premise to eliminate its begging the question. I am not saying the distinction between rational/irrational inference is the same as that between rational/irrational causation. I accept the former but not the latter. The materialist cannot accept a premise that begs the question against him, but if we can reword it in a way that is satisfactory to both dualists and materialists, then evaluation of the AFR can continue. So I make the suggestion that we reword the premise as 'no belief is rationally inferred if it is physically caused', which I can accept for the time being as not begging the question, and which I hope agrees with your intuitions. We can then investigate the meaning of 'rationally inferred', over which we probably will disagree!
SteveK, If you say the Mill is behaving rationally then you've eliminated the option of the Mill being fully explained in non-rational terms. Well, maybe. But I haven't said anything like 'the brain behaves rationally', for example. We are talking about rational inference. My position is that rational inference is a species of patterned behaviour, and patterned behaviour can be exhibited by physical systems. I gave the example of an adder circuit that behaves in accordance with an abstract pattern we call 'addition'.
David,
So I make the suggestion that we reword the premise as 'no belief is rationally inferred if it is physically caused', which I can accept for the time being as not begging the question
If you can accept this formulation and not the other then it seems you've put your finger on the crux of the disagreement. If a phrase has to be changed then, at least to you, the argument will been changed. So perhaps we should discuss the difference is between 'physically caused' and 'non-rationally caused'. Why change the argument just in order to address a non-fundamental point? (As an aside, I disagree with your paraphrasing of the premise since you left out the 'fully explained' phrase.)
Are physical causes the same as rational causes then? If not, how is the changed phrasing something a materialist can accept while rejecting the unchanged phrasing?
David,
I posted the above before reading your reply to SteveK. You explained your position clearly.
My position is that rational inference is a species of patterned behaviour, and patterned behaviour can be exhibited by physical systems. I gave the example of an adder circuit that behaves in accordance with an abstract pattern we call 'addition'.
But the 2 patterned behaviors are fundamentally different.
When we 'add' we think we are taking 2 numbers and deriving the sum (another number) according to some mental rules. An adder circuit does not think. It does not take numbers as inputs, apply some mental rules and come up with a third number.
Numbers have no extension in space, no mass, and no momentum and so cannot be affect or be affected by physical entities, forces or processes. So numbers cannot be an input to a physical system causing some physical output or even a numerical output. On the other hand our intellect can and does use numbers as inputs and outputs in the process of addition.
When we mentally assign symbols to a physical system, nothing has changed in the physical system itself. So a circuit does not suddenly start 'adding numbers' merely because we recognize we can use it to help us add. It is still reacting to voltage fluctuations according to physical laws whether we find it useful for some purpose or not.
And I thought I had made a generous offer :-) I say the original argument begs the question. At 1:50 PM you appeared to agree. This is not non-fundamental! But by all means re-insert the 'fully'. From my point of view it does no work.
David,
'no belief is rationally inferred if it is physically caused'
I'm stuck on the point I keep trying to make. It's about 2 contradicting realities that the materialist says is true - or perhaps it involves an equivocation on their part. Maybe I'm not understanding something correctly. That's very possible. Then again, maybe it's you :-)
Besides denying the AFR premise, what I know is that the materialist will claim everything in the AFR premise is material/physical.
The materialist will reword the AFR premise to something like this: "some beliefs are rationally inferred and physically caused", however in doing so they are also saying "some physical systems are not fully explained by non-rationality".
This is simple logic. If, as the materialist has said, a physical system is a rational system then it's not fully explained in non-rational terms. Now here's the contradiction / equivocation as I see it:
Because the materialist says that non-rationality isn't enough to fully explain some physical systems they cannot *also* say that a system of non-rational matter fully explains those same physical systems.
Okay, what am I missing here?
Bmiller, feel free to reply if you see something. It's bugging me.
David,
This is not non-fundamental!
What I meant was your suggestion we move on to a different area of disagreement rather than this first one which I agree is fundamental.
SteveK,
Nothing to add. I'm interested in the response too.
David,
Hopefully you will find this relevant, but, if not, then maybe it will be at least a little interesting. You said: I gave the example of an adder circuit that behaves in accordance with an abstract pattern we call 'addition'.
A first problem with the adder circuit analogy, and I think it was bmiller who pointed it out, is that the analogy is parasitic on a designer with a mind that has not yet been shown to be itself nomologically determined. With addition or counting in mind (given that counting is essentially addition), I wondered whether some other example could be given, an example which, assuming physicalism, did not depend on a designer or mind which might not be nomologically determined. And I recalled one of the only two things I remember from a plant physiology class; the professor once said that a banana (which we reasonably regard as lacking a mind) would only flower and fruit after 45 leaves had been produced. Given its mindlessness, the banana might work as an example of purely physical addition; then again, for reasons to be discussed, the banana might not end up exhibiting purely physical addition.
Anyhow, back when I was taking that class, I took the professor at his word, but now I decided to check (albeit only very very briefly) into his assertion. Of course, for the sake of this discussion, I assume that the subject banana is a wild variety rather than one which has been developed via mindful human intervention. I found this article which notes: "Generally, a banana plant produces around 40 leaves before fruiting." That generally descriptor could mean that the leaf count differs between varieties, or it could mean that different individual plants of the same variety will fruit upon the production of slightly different numbers of leaves. But is it clear and obvious that counting/addition is actually occurring here? Or is it the case that "[g]enerally ... 40 leaves" are necessary before fruiting because of purely physio-chemical conditions? This is to say that there could very well be no counting/addition occurring with regards to the banana in and of itself; it could well be that the relevant parameters are most rightly explained in physio-chemical terms without reference to leaf (or even any other) counting.
Of course, as cultivators of bananas, humans with minds that can abstract (and keep in mind that those minds have yet to be established as nomologically determined) would find leaf counting a significant pragmatic practice even if/though the plant does no counting/addition. This also gets me to wondering whether physio-chemical processes ever depend upon abstraction in order to process or whether the very notion of abstraction is just some kind of short-hand used in place of physio-chemical description. But that matter is not for this discussion, of course.
In any event, and getting back to the analogy, there remains a lack of an example of how addition/counting or any other abstract thinking process is relevant to physicalism/materialism/naturalism if that conscious thinking is epiphenomenal.
Michael,
Your banana plant example is similar to David's statement about apples.
David: "Did the physical process of apples falling from trees perform an arithmetic sum? In a sense, Yes. Suppose the apples were of equal mass and fell with equal velocity. Then the impulse delivered to the Earth was five times that of a single apple's falling."
My rebuttal to this is that rationality involves more than just the final response or output. Rationality also involves how you get to the response. Falling isn't a rational process performing an arithmetic sum. The same is true about the adding circuit. The designer of the circuit went through the rational process and the circuit is mimicking.
SteveK,
I agree. Since I was not clear enough, let me try to clarify. I do not think leaf counting is occurring with the banana; I do not think such abstract (mental) processes are effects in physicalism - at least any physicalism that insists upon nomologically determined minds/conscious thinking. I leave open the possibility of an asserted physicalism which does not insist that conscious thinking is nomologically determined and/or epiphenomenal. Whether anyone would want to call that physicalism is another matter.
SteveK,
the materialist will claim everything in the AFR premise is material/physical. Not this one! My view is that an essential aspect of rational inference is abstract, not concrete.
If, as the materialist has said, a physical system is a rational system then.... Again, not this one. I'm reluctant to apply 'rational' to objects or stuff (eg, as in non-rational matter). It's true that man is a rational animal but the sense of 'rational' here doesn't carry the same normative force as it does in 'rational inference'. The latter is an achievable ideal. Rational animals make mistakes. More reason to pursue a common understanding of 'rational inference'.
Rationality also involves how you get to the response Indeed! Let's apply that idea to inference.
The designer of the circuit went through the rational process and the circuit is mimicking. Isn't mimicry a kind of abstract correspondence, which is all I ask?
Michael,
the analogy is parasitic on a designer with a mind that has not yet been shown to be itself nomologically determined. I'm not sure what role the demonstrated nomological determinedness of the designer's mind plays in this claim, but I'll assume that the claim implies that the analogy is parasitic on a designer with a mind. But I'm not sure that this derived claim is true. There is an ontological issue here and a genealogical one. I grant that the adder is an artifact. Nevertheless, even if we had no idea as to its origin, we can see that its behaviour conforms to the abstraction we call 'addition', especially if its structure can be reverse-engineered.
I too doubt that fruiting in banana plants involves any process that instantiates addition. Non-linear biochemical dynamics that attracts to 40-ish is much more likely an explanation.
Could you say a bit more about nomological determinateness and its role as you see it in this question?
David,
Nevertheless, even if we had no idea as to its origin, we can see that its behaviour conforms to the abstraction we call 'addition', especially if its structure can be reverse-engineered.
I disagree. As I mentioned above, when we 'add' we 'add' numbers. Numbers are not physical things. Whatever the circuit is doing it is not doing anything with numbers.
David,
My view is that an essential aspect of rational inference is abstract, not concrete.
Isn't the standard materialist view that *all things* are the result of material interactions? On that definition the abstract rational inference is only the result, or conclusion, of the material interaction. Because the abstract rational inference isn't a material interaction it's not a mental process that we would call thinking. The thinking is done at the level of material interaction.
I honestly didn't know David was a materialist. So much for me being observant, lol. Old age is my excuse.
Or non-rational causes :-)
BM, Or non-rational causes :-) Aha! Very Good! Numbers are not physical things. Whatever the circuit is doing it is not doing anything with numbers. OK. Would you accept that it is doing something with representations of numbers?
DB: My view is that an essential aspect of rational inference is abstract, not concrete. Let me expand on this a bit. Rational inference appears to be a process whereby a new belief arises from existing beliefs. We don't really know what a belief is. We think of a belief as something inside us, internal rather than external, classified as mental rather than physical. But whatever a belief really is we can express it in language, in a sentence, and convey it to others. And so a belief enters the physical external world. Note that a sentence is already an abstract thing because it can take the concrete form of spoken sounds or written squiggles. We can recognise that sounds and squiggles can represent the one sentence. Some people will go further and abstract from particular languages and writing systems to so-called propositions, but I will stop at sentences. Now, rational inference seems to me to be an 'operation' on patterns of sentences. So we have ascended another level of abstraction. 'All men are mortal' abstracts to 'All A are B'. 'Socrates is a man' abstracts to 'N is A'. Rational inference matches these two patterns together and produces the new pattern 'N is B', and then dropping down a level of abstraction from sentence patterns back to sentences we get to 'Socrates is mortal'. Note that inference is independent of the meanings of the categorematic words 'Socrates', 'Man', and 'mortal'. But not independent of the meaning of the syncategorematic word 'all'. If we replace 'all' by 'some', this inference does not proceed. Finally, the new sentence is de-expressed, as it were, to a new belief, just as if the sentence were told us by a trusted source. So that is a theory of rational inference of belief, in outline. The final thing to say for me as a materialist is that I cannot find a reason to think that any of these steps cannot be realised physically. Do enlighten me :-)
David,
The final thing to say for me as a materialist is that I cannot find a reason to think that any of these steps cannot be realised physically. Do enlighten me :-)
I raised some issues in my previous comments and bmiller has too. Some quick points...
1) If these steps are happening physically then it's not fully explained by non-rational things. Not a problem is you don't ever say that sort of thing.
2) These "essential abstractions" cannot be physical. You said they weren't "concrete" which I interpreted similarly. That makes the process *not* entirely physical. Thoughts?
3) If it's entirely physical then it seems to me that logic must be a natural law of some kind. The physical process is guided, via forces, to follow a logical path. Is this your view? Why hasn't a physicist ever attempted to put this into an equation?
David
More thoughts on the assumption that everything is physical:
A physical abstraction has no inherent meaning and no semantic relationship to other physical things. It can be an physical effect of a physical process, sure, but it cannot have meaning or even "point to" something beyond itself or what caused it.
The physical abstraction of "Socrates" has no meaning of Socrates any more than a shadow does. A shadow of Socrates doesn't capture anything about Socrates being a man or a cardboard cutout. Also, there's no relationship to the other different abstractions. They are all independent with no physical law that can manipulate them into forming the final abstraction "Socrates is mortal" while keeping other physical abstractions away.
Physical abstractions of Socrates, a dead man, the sky, trees, Fred and other abstract thoughts all bouncing around in your brain come together and form the final physical abstraction "Socrates is mortal"? Why not "Socrates is the sky" or "Fred and Socrates are dead"? Explain the physics that creates this very controlled and specific order?
David,
You asked: Could you say a bit more about nomological determinateness and its role as you see it in this question?
First let me address or clear up the matter regarding something else you said: I'm not sure what role the demonstrated nomological determinedness of the designer's mind plays
I expect you grant that it is common to have the experience of conscious thinking seeming to be an active - an actual - cause rather than a mere dead-end effect, an effect which itself does not cause but which, at most, is some sort of inert product of an uninterruptible wholly physical process wherein conscious thinking and consciousness are only epiphenomenal and do not contribute to the continuation of the causal-chain physical process. Phenomenologically, conscious thinking is not a mere inevitable by-product of physics; hence, phenomenologically, a designer is one whose mental activity is not a mere by-product result; therefore, phenomenologically, the designer's mind is NOT presumed to be or experienced as being nomologically determined (even though the mind is constrained or contextual). The point here is that the adder circuit entails a designer, and, phenomenologically, that mind is presumed to effect what it does owing not to nomological necessity but, rather, owing to that mind facing a non-determinate situation in which the mind can design the circuit one way or another or not design the circuit (whether by choice or by failure). I brought up the banana issue as a way to try to come up with a situation which would better avoid the phenomenological experience of non-determinateness.
Nomological determinateness is a slightly more brief expression of nomologically necessary determinateness. As noted here, in the Varieties section, "Nomological determinism is the most common form of causal determinism. It is the notion that the past and the present dictate the future entirely and necessarily by rigid natural laws, that every occurrence results inevitably from prior events." Likewise, as noted in the Nomological determinism section of this article, "Probably the most common form of causal determinism is nomological determinism (sometimes also referred to as hard determinism, as physical determinism, or as metaphysical determinism) ... the view that all future states and events of the universe are dictated by prior states and events of the universe together with the prevailing laws of nature." These descriptions are reiterated here: "Nomological determinism ... is the notion that the past and the present dictate the future entirely and necessarily by rigid natural laws and that every occurrence inevitably results from prior events."
In effect, nomological determinism (maintaining the terminology from the previous paragraph) denies that there are actual possibilities; there are only necessities. Accordingly, there are no actual physical or metaphysical ifs.
So, when SteveK asks whether "logic must be a natural law of some kind" and whether the "physical process is guided, via forces, to follow a logical path", the answer would have to be a resounding NO inasmuch as actual possibilities which are denied by nomological determinism are necessary to logic. Without actual metaphysical or physical possibilities - the ifs - there is only description and there is no actual logical path.
To be continued ...
Continuing ...
In case the previous part of the posting does not satisfy the "say a bit more" request, let me refer back to an earlier posting. You had said, "[Lewis] certainly seems to think that rational inference depends on some extra-physical human capacity", and I replied, "But does it have to be 'extra-physical' if physicalism is the case but utter nomological determinateness is not? Hopefully, this sheds some light on why I think the issue is about - or includes - the matter of conscious thinking and whether or not it is epiphenomenal." Rational or logical inference does not in itself appear sufficient to dispense with physicalism if there can be a physicalism without nomologically necessary utter determinateness. Another way of considering this issue is in terms of whether physicalism can accommodate consciousness as actually non-epiphenomenal. That could be worthwhile, foundational common ground.
David,
Thanks for the longer explanation and how your thoughts regarding materialism seem different than others.
Here are my thoughts:
OK. Would you accept that it is doing something with representations of numbers?
What the machine is 'doing' is changing physical states in response to external stimuli. The machine has no intent to represent anything. So, no I cannot attribute that activity to the machine. If anything is doing any 'representation' wrt numbers and physical states it is a human mind, not the machine.
The final thing to say for me as a materialist is that I cannot find a reason to think that any of these steps cannot be realised physically.
But all of those steps were performed mentally by a human mind aside from the physical part of writing/reading or talking/listening. Machines don't have beliefs, draw conclusions from premises, operate with or on numbers etc.
And so a belief enters the physical external world. Note that a sentence is already an abstract thing because it can take the concrete form of spoken sounds or written squiggles. We can recognise that sounds and squiggles can represent the one sentence.
But a person's belief does not enter the physical external world via sentences whether written or spoken. Squiggles and sounds may have, but not a belief. The belief has not been transformed from being an internal immaterial subject of the human intellect to being physically present in ink on paper. If it were, then a belief could be smeared, erased or burned. I suggest the sentence is a sign pointing to a belief and a sign can be either a physical object or a mental concept but it is not the thing being signified.
Is the process of process of signification itself just an instance of rational inference? If so, then explaining rational inference in terms of itself would seem circular.
On second thought we need rational inference to investigate anything.
Perhaps it helps us to use signification while rationally inferring.
Thank you for your thoughts, Gentlemen. I will make some responses and perhaps some alterations today. I am already regretting using the term 'abstraction'.
Thanks, Michael, I understand you better now. I agree that we experience ourselves as free and undetermined. I ascribe this to epistemic possibility rather than ontological. We just don't sense enough about ourselves to see the rails on which we run.
Physics tells us that the world is indeterminate at the smallest scales but at the scale at which we sense it the indeterminacies mostly average out to a determinate picture. They can, however, be amplified up to sensible scales. Geiger counters, for example. But I assume that our brains can be understood through classical physics and chemistry. (There have been claims that migratory birds sense the Earth's magnetic field through quantum phenomena but the usual objection to this is that brains are too warm and 'noisy' to amplify successfully these effects to the scale of nervous switchings.) I suggest that it is the classical determinateness that guarantees our ability to make rational inferences.
You say actual (metaphysical, ontological) possibilities are necessary to logic. But consider the static, possibility-less, world of a painting of coloured blobs. Logic still applies to the relations between the blobs. If a blob is blue then it has a green blob to its right, say. This is only description, yet it is description in logical terms.
BM,
What the machine is 'doing' is changing physical states in response to external stimuli. And we are doing more?
The machine has no intent to represent anything. Nor does a map. Yet it does represent the topography. Would the map in the satnav of a self-driving car (an auto-auto) not be representing the topography to the car? Will you say that it is just digital bits? But the bits have many layers of structure, of patterning, to which the car is sensitive---or if that is too anthropomorphic---to which the car reacts. This is the idea that we have discussed before that something is representational if it can guide an entity in its navigating the world. Literally, in this case.
So, no I cannot attribute that activity to the machine. If anything is doing any 'representation' wrt numbers and physical states it is a human mind, not the machine. OK. The 'derived intentionality' issue.
Machines don't have beliefs, draw conclusions from premises, operate with or on numbers etc. Not in the way we do, perhaps. But that doesn't rule out these phenomena being physical in us. That's the present question.
But a person's belief does not enter the physical external world via sentences whether written or spoken. You are right. This is philosophy not poetry. Perhaps I should have said a representation of a belief. But my original has a metaphorical truth. The only way minds have of copying belief from one to another is via speech or writing. We discussed this not so long ago, I think. And I would add that it's primarily through sentences that our beliefs are apparent to ourselves. Or rather our conscious selves perhaps. A belief that a black dog is dangerous, say, in conjunction with the appearance of said dog, can affect our bodies without conscious rehearsal of the belief. And then we sense a change in our bodies and become aware of the belief, perhaps in a secondary way.
Is the process of process of signification itself just an instance of rational inference? No, not in the narrow
understanding I'm working with. The meanings of the categorematic terms in an inference are irrelevant, or so I claim.
David,
What the machine is 'doing' is changing physical states in response to external stimuli. And we are doing more?
We are doing something different. We are adding numbers.
Not in the way we do, perhaps. But that doesn't rule out these phenomena being physical in us. That's the present question.
Numbers have no extension in space, no mass, and no momentum and so cannot be affect or be affected by physical entities, forces or processes. So the way we manipulate numbers cannot be physical either.
Regarding navigation software. It's basically no different from player piano rolls.. I don't believe player pianos have human intellects either just because someone loads it up and hits the start button.
BM,
So the way we manipulate numbers cannot be physical either. Turn this around: we don't manipulate numbers at all. We manipulate representations of numbers. Our problem is that we mainly know numbers through their representations. So we tend to confuse the two. Compare beliefs and their representations. In my case, guilty as charged. And we can also grasp and represent the patternednesses within the numbers. As to what they are...Discussion for another day!
piano rolls. It wasn't a question of intellect possession. It was a question of representation. There is a clear sense in which piano rolls represent or simply encode music.
Hi David,
Our problem is that we mainly know numbers through their representations.
How do you know that we know numbers mainly through their representation and not directly? That is wihout begging the question that only the physical exists.
Even if we can signify some physical state to represent some number it still must mean that there is some number there to represent in the first place. And in the case of a number, it would mean that a non-physical entity exists.
It wasn't a question of intellect possession.
But we've been asked to believe machines are doing the same things humans do such as 'adding'. My position is that they are not doing the same things as humans. Something has to be in possession of an intellect in order to claim something or the other is adding. My claim is that it is not the machine or software running it.
David,
There is a clear sense in which piano rolls represent or simply encode music.
This is mimicry. The piano doesn't grasp the melody or the timing signature or anything related to music itself. It doesn't know that it's playing music so it cannot actually be *representing* music. I agree with bmiller, the possession of an intellect is required.
We manipulate representations of numbers. Our problem is that we mainly know numbers through their representations.
"Mainly know" or "fully know"? It's an important distinction. "Mainly" implies that there is a pathway to know numbers directly. Do I have knowledge of the external world around me directly, or do I only know the representation?
The only way a piano roll can represent music is if an intellect is involved in the process. Rain drops falling on a drum kit mimics music. An intellect makes music.
David,
After wondering whether to bother with a response which would not be expected to effect (certainly) immediate resolution of differences, I have decided to offer the following in the belief that it is ontologically possible (HaHa) to be found eventually to be at least a little interesting.
You said, "I agree that we experience ourselves as free and undetermined. I ascribe this to epistemic possibility rather than ontological." Thinking in terms of categories such as epistemic and ontological can certainly be useful; I see them as akin to isolating or controlling for variables. However, these categories must fit back together - they must be compatible - just as it is the case that variables that have been isolated certainly fit back together (this might just be a personal preference, but I am inclined to thinking beyond categories). Retrospective analysis certainly expects to eventually eliminate the epistemic possibilities which constitute a particular ignorance in order to arrive at (let us say) ontological fact. The reasonableness of such an expectation derives from the even more reasonable expectation that there is no retrospective ontological non-determinateness. Eliminating epistemic possibilities in such a manner says nothing about - and does not depend upon - whether what is now the past, before it was past, was devoid of ontological possibilities/alternatives.
That being said, there is a sort of determinateness which is not incompatible with actual possibilities, and that determinateness would be a context consisting of alternative, conflicting possibilities. For example, the context consisting of the possibility that I do write tomorrow and the possibility that I do not write tomorrow would be a determinate context consisting of and compatible with actual alternative possibilities. Whereas retrospective epistemic possibilities can be reasonably presumed to indicate ignorance rather than non-epistemic, ontological non-determinateness, it does not follow necessarily from the ontological determinateness of the past that the future is determinate in a like manner.
You also remarked: "I suggest that it is the classical determinateness that guarantees our ability to make rational inferences." Ah, but that is very different than the notion that classical determinateness makes necessary our rational inferences. Determinateness is certainly regarded as necessary for the orderliness of the mindless and the inanimate to be observed and predicted, but that does not speak to the experience of conscious thinking as non-epiphenomenal. No one denies that rationality either depends on or presumes or expects orderliness or even determinate contexts as mentioned above; however, what is never demonstrated is that nomologically necessary utter determinateness is the case, and what is also never demonstrated is that rationality (including legitimate inference) is identical to or dependent upon nomologically necessary utter determinateness.
Finally, you also said, "... consider the static, possibility-less, world of a painting of coloured blobs. Logic still applies to the relations between the blobs. If a blob is blue then it has a green blob to its right, say. This is only description, yet it is description in logical terms." Well, I would never use logic and logical in that way. But that is not a big deal or a problem at all, because, as I believe I might have indicated some time ago, I do not believe any expression is necessary; I expect expressions to be alternatively expressible.
SteveK, Let me apologise. My 'rational inference is abstract' was a bit quick and glib. What I meant was that it is dependent on finding, matching, and manipulating patterns, and patterns within patterns. Brains are pretty good at this. But when we find patterns we are paying attention to some aspects and ignoring others, so we are abstracting. But the result of the abstraction has to be another concrete representational thing (sorry, BM) that can be further worked on. And so it goes on. The justification for saying this is all possible physically is (a) we can express rules of inference as pattern transformation, and (b) there are computer programs that implement these transformations in the course of finding and proving theorems.
Your point that logic must be a natural law of some kind is interesting. Given a language with terms for individuals and concepts they can fall under I would say that logic encapsulates the meanings of 'some', 'all', 'and', 'or', 'if-then', etc, which seem to arise naturally.
A physical abstraction has no inherent meaning and no semantic relationship to other physical things. I think that's right. It just represents a pattern.
They are all independent with no physical law that can manipulate them into forming the final abstraction The patterns we are talking about represent sentences by means of symbols. There is no problem making a physical system that juggles the symbols according to certain rules to make a new pattern.
How do the appropriate thoughts or beliefs come together in a controlled or specific order? A good question. The is the so-called 'frame problem' that bedevilled old-fashioned AI. Perhaps these days massive parallelism or a neural network architecture could find the right antecedents to an inference quickly. But we humans don't always find the right logical path ourselves, and we sometimes make invalid inferences.
David,
We can physically manipulate physical objects. Objects with extension in space, mass, velocity and such. How do we physically manipulate a 'pattern'? Does it have any of those 3 attributes?
If a pattern is the shape of an object for instance how does that get into the brain? Wouldn't it get crowded in there if we were considering a cow?
Bmiller
I can see how an external object creates the effect of a brain pattern. No different in my mind than an object creating a shadow, air movement pattern or an electrical disturbance of some kind. However, all of those representative symbols are empty in the sense that they don't represent much of anything about the object itself.
None of those things can tell us if Socrates is a man because there's nothing around - no "you" - to interpret the symbols correctly. Artifacts can't be known to be artifacts. Socrates without his legs is a different representative symbol that Socrates with his legs. Is that the same thing or a different thing (essence)?
BM and SteveK,
How do you know that we know numbers mainly through their representation and not directly? From our childhood learning experience. After learning to chant the number line we are introduced to the symbols for single digit numbers and then two digit numbers and then the process of adding. This is knowing them through their representations as sequences of decimal digits. We tend to fall into thinking of numbers as objects. But they might as well be thought of geometrically as places along the number line, or movements back and forth along it. I think that is the main way we come to know them. A secondary way is through a mental model or image of some sort. I think of the natural numbers as a row of fence posts stretching away to infinity. A sort of model of the Peano axioms. What would it mean to know them directly? Through the senses, perhaps? Surely not.
But we've been asked to believe machines are doing the same things humans do such as 'adding'. Ah, I think I see why we are talking past each other. I was trained in maths. We math weirdos think of any function mapping a pair of things to a third thing that's both associative and commutative as an addition. We usually demand a 'zero' thing, something that when added to a thing leaves it unchanged. It's just a pattern. Lots of kinds of things can be added, not just numbers. Obviously it takes an intellect to see if the things and the function conform to the pattern. But we don't ask how the function 'works'. It's just a given or observed relation between pairs of things and third things. It's rather superficial, really.
The piano doesn't grasp the melody or the timing signature or anything related to music itself. It doesn't know that it's playing music so it cannot actually be *representing* music Where does this idea come from? This is surely way too demanding a requirement for representation. It would deny that photos, portraits, films, recordings, sheet music, etc, were representations. Besides, my claim is that it is the piano roll, CD, or whatever that counts as the representation not the piano or CD player. The latter are more like artificial sense organs or transducers.
David,
This is knowing them through their representations as sequences of decimal digits.
Maybe I misread or misunderstood initially that you meant that we only understood numbers as their representations. But then this statement says that we know them through their representations. If we know them then we know them. Then again your further remarks leave me in confusion what you think. Maybe we should start with this definition from Wikipedia:
A number is a mathematical object used to count, measure, and label.
Is this a good definition? If so then it seems they have no physical traits. How can we know something that doesn't have any physical traits from the perspective of a materialist?
Ah, I think I see why we are talking past each other. I was trained in maths.
I have been trained in digital design and test. So your example of a digital adder means a series of gates connected together in a certain manner for a specific purpose to me. Then maybe that is where we are talking past each other, but I find it odd regarding our relative positions.
The building blocks you are using from a mathematics perspective are completely detached from physical reality. There are no functions, mappings, associative or commutative laws that physically occupy space and time in the physical world. On the other hand, I have physical voltages, transistors, conductive paths and such that I can actually point to and measure with instruments. Yet it is you arguing that your own world doesn't really exist and mine does. It seems our positions should be reversed!
David,
Besides, my claim is that it is the piano roll, CD, or whatever that counts as the representation not the piano or CD player.
You may have missed where I said "The only way a piano roll can represent music is if an intellect is involved in the process". I've discussed this already so I won't repeat myself.
You seemed to be in agreement with me here:
Me: "A physical abstraction has no inherent meaning and no semantic relationship to other physical things."
You: "I think that's right. It just represents a pattern."
SteveK,
I can see how an external object creates the effect of a brain pattern. No different in my mind than an object creating a shadow, air movement pattern or an electrical disturbance of some kind.
But I think that is skipping steps.
The object presumably possesses a 'pattern' of some sort that somehow gets into the brain. What is that physically? If the pattern is the shape common to all cows, then how does something of that size get into something the size of a brain? It may seem like a silly thing to say, but I bring it up to illustrate that no one thinks a pattern is something that has physical attributes that ends up physically in the brain.
I don't see how one can even start to talk about patterns without having first assumed that patterns are immaterial things. Immaterial things are not available for materialists I presume.
bmiller,
no one thinks a pattern is something that has physical attributes that ends up physically in the brain.
I don't understand. Brain pattern tests record the waveforms and various electrical activity in the brain. Those are patterns, yes? I know you are aware of this so you must be intending to say something else and I am failing to grasp your point.
SteveK,
Brain pattern tests record the waveforms and various electrical activity in the brain. Those are patterns, yes?
The patterns of electrical activity from the brain are not the patterns of cow shapes. If we are physically manipulating cow shapes in our brain then the cow shape is in our brain. That was my point.
But that doesn't mean that physical things don't exhibit patterns that intellects can recognize and abstract from those physical objects. It's just that patterns per se have no physical being apart from what they are patterns of. So I agree that we manipulate patterns, but the patterns apart from the objects they are patterns of. Since they have no physical being apart from their physical object we must be manipulating immaterial objects.
bmiller,
I understand. I was willing to accept that the patterns got into the brain somehow so that we could focus on what I think are bigger problems for the materialist. But, yes, I agree that looking at a cow doesn't produce a cow-shaped pattern in the brain. I've brought up the issue of meaningless patterns to David and it appears he agrees, but maybe not.
Me: "A physical abstraction has no inherent meaning and no semantic relationship to other physical things."
David: "I think that's right. It just represents a pattern."
The patterns in the brain need to be interpreted. You look at an object and now the brain somehow has a pattern. What does the pattern mean? No clue. Socrates puts on a hat and coat and the brain has new pattern. It compares it to the first pattern where no hat and coat were involved and the result is the same - no clue. The brain has no idea what these things are because all the brain is working with are the physical patterns. The brain doesn't know essential properties from accidental properties either (I use the term artifact before by mistake).
Seems like we are looking at different aspects.
David will be pleased that he has so much material to work with :-)
Our conversation is getting into issues related to Searle's 'Chinese Room'. I've always thought that his thought experiment, actually a similar one, could be performed as an experiment using humans instead of a machine.
1) Put humans in a room. Let them observe everything in the room to generate 'brain patterns'.
2) Give them a box of symbols. These are the patterns that represent what they observed.
3) Have them arrange the symbols according to a set of pre-established rules.
4) Ask them if they know what any of the symbols mean, which object each one refers to or what the arrangement intends to communicate. (hint: they wouldn't know anything)
If conscious humans can't do this neither can a machine.
Apologies, gents, was away all day yesterday and unable to comment.
Firstly, SteveK's objection that any representation requires an intellect to interpret it. I can see why one might say this: all the common cases of representation involve things for we humans to make sense of. Yet it seems quite reasonable to me to say that the image of a fly on a frog's retina represents the fly to the frog. Or to the frog's brain. But maybe I should use a different word altogether. The best I can come up with is 'encoding' or perhaps 'model'. That feels rather more mechanical. And it seems to work for piano rolls and CDs and digital maps for self-driving cars.
The problem we are grappling with is that I have been trying to account for inference in terms of patterns yet it would seem that a physical thing would have no way of interacting with something abstract like a pattern. So what is the next best thing? Answer in next comment.
The next best thing has to be a representation of the abstraction. Or an encoding or model. We can't interact with numbers directly, but we can interact with strings of decimal or binary or whatever base digits. The marks '2', '3' in that order are a lot more concrete than the number twenty-three, I'd say. I can just about add up a pair of two-digit decimal numbers in my head and I tend to resort to mentally picturing what I'd write down on paper. Most additions with more digits than this overload my short term memory and I am obliged to write them down on paper or use a calculator.
The marks *AB, nA → nB in that order concretely model a family of valid Barbara style inferences. The marks can be understood as denoting an inference rule. Roughly, given two sentences one of which matches *AB and the other which matches nA, validly infer the sentence nB. A sentence matches *AB if it begins 'all' or 'every', followed by a concept word A and another concept word B. In order to match nA the other sentence must be a name word n followed by the same concept word A as in the former sentence. The inferred sentence is the name word n followed by the concept word B. The important point here is that this is wholly mechanical. With some details filled in about representations or encodings or modellings of words and sentences our marks could be compiled into a program that returns inferences given sentence pairs. Another important point is that inference is thus a syntactic business, not a semantic one. It matters not what the words mean, and they mean nothing to the mechanical manipulator.
This is not to say that this is how we infer. It just says that inference could be mechanical and hence physical.
SteveK says, The brain has no idea what these things are because all the brain is working with are the physical patterns. Absolutely. But a brain that worked this way and gave intelligible responses might give the impression to an observer that it knew what it was talking about. I think our conscious selves are observers of our own responses, so I think my own brain knows what it's talking about. Consciousness, though, I think a complete mystery. My arguments suggest that a language-using robotic entity capable of modelling a limited part of its environment, interacting with it, making inferences, and so on, is possible. It would have no sense of self. It would be forever sleepwalking, as it were. Nobody would be at home.
David,
Since a purely physical thing cannot possibly interact with an immaterial pattern, then how can you, as a purely physical thing, create a physical representation, or even any representation at all, of a purely immaterial pattern?
It also seems to me that representation entails abstraction unless the representation is a complete copy in all respects. So the purely physical you, would have to somehow manipulate an immaterial abstraction (the pattern of inference) to produce a different abstraction (the proposed physical model) in order to physically begin to build this machine. It seems to me that the 'next best thing' just repeats the original problem twice.
David
What bmiller just said is a very big part of the problem, and what I am attempting to do is highlight other problems. It's not like I expect anyone to have all of this figured out. Nobody does. What I expect is for an explanation to make rational sense on it's own terms.
Absolutely. But a brain that worked this way and gave intelligible responses might give the impression to an observer that it knew what it was talking about.
You just said "Absolutely" that the brain has no idea so how can a different brain in another observer know? Each brain is functioning on the same principles. You aren't being consistent.
Yet it seems quite reasonable to me to say that the image of a fly on a frog's retina represents the fly to the frog.
This is where bmiller and I intersect on the same problem. The essence of "fly-ness" is immaterial. I discussed this previously when I talked about universals vs. particulars and when discussing Socrates shadow and Socrates with/without a hat/coat. Humanness is not captured in a photograph. The photograph doesn't inform the brain what physical aspects belong to the human (essential) and which do not (accidental).
BM, You are now asking the historical question, How did something get to be the way it is? rather than the ontological question, What is it? How does the representation of the inference rule (in us, not robots built by us) come about? I think a materialist has to answer, Through some sort of Darwinian trial and error process. The lower storeys might be built by genetic variation and natural selection. The upper storeys by the process of language acquisition, which is also a kind of trial and error. We should find it significant that natural language has deductive inference built into it in such a simple way, reason and language tightly bound together, as it were. I can't say anything beyond this. What I have argued so far exposes a gap in the AFR, in my view. In effect I am saying, Why can't things be the way I describe? and the onus falls on the defender of the AFR to explain why, in order that his explanation remain the only candidate. And he too has an historical question to answer, for that matter.
SteveK, Let me step back a bit. You said that the patterns in the brain need to be interpreted. I'm not sure that's the right word. It suggests 'made sense of' in some way, and I don't think that is required (or possible). A rule of inference pattern representation rather has to acted on, executed in some way, and this is not beyond the capabilities of a machine. I'm also not sure that considerations in respect of consciousness are relevant here. I think a rational, language using machine with a kind of 'blind spot' where its self or I should be is possible. It would have no way of reflecting on its own mental activities, as it were. Language use and inference could be independent of consciousness.
I'm afraid I didn't follow what you said regarding essences, universals, and particulars. How do these ideas interact with your conception of representation? They don't figure in mine at all.
David,
You are now asking the historical question, How did something get to be the way it is? rather than the ontological question, What is it?
I am asking 'what does it do?' because that will tell me what it is. If it doesn't do anything at all it is one sort of thing and if it grows teeth and chases gazelles it is another sort of thing. That is how I can tell a statue of a lion from a real lion. Likewise a purely material thing will not be able to do immaterial things.
In effect I am saying, Why can't things be the way I describe?
Because it appears you are ascribing immaterial operations to purely material things within a theoretical framework that disallows immaterial operations.
I may have misunderstood. You used the word 'create' so I took you to be asking how something comes about. Namely, how does a purely physical thing (me), incapable of interaction with an immaterial pattern, bring about a representation of that immaterial pattern within itself? Have I understood? My answer is by trial and error, aided by another purely physical thing that has already acquired a whole host of similar representations, namely a parent teaching me to speak and infer in a natural language.
Yes, I think you misunderstood.
It seems we agree that a purely physical being is incapable of interaction with an immaterial pattern. I have assumed that for a materialist, immaterial things do not exist since all that exists are material things. If there is nothing to represent, it seems there can be no representation of it. Maybe not so for you.
If the materialist allows that immaterial things exist but he cannot interact with them, then how can he be aware that they exist at all? I don't think it does any good to say he interacts with a 'representation' of this immaterial thing because the one making the representation would have to be aware of the immaterial thing in order to make (create) the representation. If one says he is aware that the immaterial thing exists because he has encountered a 'representation' of it we are arguing in a circle.
Additionally in order to make the representation one would have to perform the immaterial act of abstraction from the immaterial thing of the immaterial feature or features one wants to represent physically.
The same goes for the one who presumably gains understanding from interacting with this physical representation only in reverse. Why would he think it was representation of anything rather than just the physical thing itself?
Suppose I have four tokens on my desktop arranged in a square. Now, as a materialist or nominalist or whatever I don't want to say that some immaterial object, a so-called 'square', also exists. Nevertheless, the concrete world of material objects in space and time allows for spatial relations between objects. They are not of necessity all in the one place. There may be no 'immaterial objects' called squares yet our tokens are in a square in reality. Language forces me to use these abstract nouns. I can get only so far with adverbs---arranged squarely, say. You will then say, If there is nothing to represent, it seems there can be no representation of it. But that cannot be quite right. There is no square material object present, and therefore no square material object to be represented. Yet the squareness of the tokens' spatial arrangement is physically representable, or so I say.
You ask, If the materialist allows that immaterial things exist but he cannot interact with them, then how can he be aware that they exist at all? I don't concede the antecedent. I don't say that there is a physical realm of space and time filled with material objects, and another immaterial realm where their spatial relationships hang out, but to which I as a material object have no access. Rather I say that the spatial relationships are intrinsic to the space, time, material object system and that we can think them by means of physical representation within ourselves.
You say, the one making the representation would have to be aware of the immaterial thing in order to make (create) the representation. We are aware of the spatial relatedness between the tokens. This is given in vision---four distinct spots on the retina---and in proprioception---sensing that our arms and fingers have to be in different places in order to touch the tokens. The next bit is tricky. I have to say something like this. The possibility of representing squareness comes about through some trial and error process as we learn our language. We don't know what the word 'square' means until we have been shown examples and maybe the fourness and equal sidedness and right-angledness made explicit. So more basic pattern representations must be already in place, or at least developing, so that they can be connected into the emerging 'squareness' nexus. This is awfully hand-wavy but it does pay respect to the fact that language and understanding have to be learnt over a period of time by exposure to speech and examples. With this picture in place what you said starts to look backwards. I say we get the capacity to represent first, through the learning process, and only then can we be aware of the 'immaterial thing', the squareness in the arrangement of the tokens, that it can represent. I hope this makes at least some sense!
David,
I wouldn't say an arrangement of tokens in the shape of a square requires the additional of an immaterial object either. But I would say that the shape has no physical existence apart from the objects. I think you agree.
But that cannot be quite right. There is no square material object present, and therefore no square material object to be represented. Yet the squareness of the tokens' spatial arrangement is physically representable, or so I say.
But I agree that there is a square material object present and that squareness is represented in that (those) objects. But 'squareness' detached from physical objects is not a physically existent thing. When you say that you can look at the tokens and end up drawing a square on a piece of paper I assume you have somehow separated the arrangement from the material objects or else you would have a 1 meter square arrangement of tokens in your head. Squareness is immaterial. Absent the immaterial mental capability to grasp immaterial 'squareness' all that would be physically present is an arrangement of tokens and ink on paper. 2 totally different physical things.
I agree with you that there are not 2 distinct realms.
With this picture in place what you said starts to look backwards.
But if I look at that picture it seems that you've conceded that 'squareness' exists. The fact that it can be represented abstractly by something physical is beside the point if this immaterial thing is said to exist and is knowable.
Objects arranged in a physical shape can be labeled a "square", however a larger amount of them arranged in a physical shape cannot also be labeled a "square" because it is a different physical shape. As bmiller said, squareness is immaterial so there's no ability for a physical system to grasp or to represent it.
The physical objects are not actually arranged in a square since they don't fit the definition of a square exactly. Something will be physically off and the physical system has no way to capture that difference via physical representation. No amount of trial and error can find something physical that doesn't exist physically.
The physical objects can be observed and therefore represented from various physical positions. Which is the correct representation or are all of them correct? The physical system has no idea. If you don't know what I'm talking about, David's physical arrangement represents a "square" from one position, a "rectangle" from another, and a "parallelogram" from another. There are countless physical representations.
I'm using scare quotes because shape concepts are not physical.
SteveK,
Yes, thanks, those are aspects I neglected to mention.
Gentlemen, some things you say in your comments on my 4:53 AM piece suggest to me that you both think I am putting forward a 'picture' theory of mental representation. This could not be more wrong. I think picture theories are hopeless.
I don't want to get into the ancient nominalist versus platonist debate about the ontology of properties or abstractions. I think a good theory in this area would explain why this debate has gone on for millennia. The idea is that we can form representations which do not represent, or at any rate do not represent objects. Example. This is a picture, yes, but I'm asking you to translate the concept of a non-representing representation into another representational mode, such as the one I claim operates in our brains and which generates representations in the course of our learning how to use a word like 'square'. What a representation does, in this context, is pattern recognition associated to a word. Thus we can learn to recognise a man, a tree, a square, and associate these recognisings with 'man', 'tree', and 'square', but there need be no abstract or immaterial things 'man-ness', 'tree-ness', or 'square-ness'. In particular, there is no perfect or exact or ideal square. Every thing is more or less square, mostly less. Also, I really should add that 'pattern recognition' here does not mean some kind of matching against an internally held image. Obviously, to learn the meaning of 'square' we have to be visually and proprioceptively presented with lots of instances of things we are told are 'square', and hopefully are square, in various sizes, orientations, distances, etc, practise drawing squares, cutting them out of paper, and so on. These experiences are individually forgotten but leave behind a physical recogniser, a representation.
In the light of this, does 4:53 AM make more sense?
David
I'm not proposing a picture theory of mental representation. There are no mental images where the brain forms a "picture image" of the frog or the objects spatially arranged in the form of a square. My comments are directed at a generic physical brain pattern in whatever physical form you want to propose. I don't see anything in your 3:43pm comment that would change any of my objections.
David,
OK. It was mainly the fact that you brought up the squareness of the arrangement of tokens that I formed my response around that particular pattern. But I don't think it changes the problem if we express the representation in a word rather than drawing a picture. We still have to acquire something external to ourselves, internalize it and express that something as a representation.
What we acquire is not a man or a square or anything else we can sense. If you say it is a representation of a man or a square you have not explained how that representation comes about, what it entails, and how we interact with it.
What a representation does, in this context, is pattern recognition associated to a word.
How does a representation do pattern recognition? Are you using representation as a verb or a noun?
Thus we can learn to recognise a man, a tree, a square, and associate these recognisings with 'man', 'tree', and 'square', but there need be no abstract or immaterial things 'man-ness', 'tree-ness', or 'square-ness'.
I think you are saying that if you can program a computer to digitize images, abstract away things like actual physical sizes, reduce 3 dimensions to 2 dimensions at the correct orientation, smell, texture etc and compare enough pictures of squares, there will be a series of bits in memory that represent the common essential leftovers and one can call squareness. Right?
This just seems like the adder argument again. The machine is not a rational agent, cannot know anything, cannot judge truth or falsity and cannot form any beliefs. It may help you find patterns in a signal as a tool, but only in the same sense that using fingers while counting helps you count or using a bright light and magnifying glass while inspecting an insect. I don't understand how this somehow means that a belief rationally inferred can also be fully explained in terms of non-rational causes.
*I hope asking you to translate the concept of a non-representing representation into another representational mode, isn't key to your argument because the wording makes my head hurt.
David,
Would you say that that brain is operating in way that is similar to Searles Chinese Room that I mentioned previously? If it's different, can you explain the differences? That would help me.
SteveK,
I know you asked David, but here is my take:
1) Put humans in a room. Let them observe everything in the room to generate 'brain patterns'.
2) Give them a box of symbols. These are the patterns that represent what they observed.
3) Have them arrange the symbols according to a set of pre-established rules.
4) Ask them if they know what any of the symbols mean, which object each one refers to or what the arrangement intends to communicate. (hint: they wouldn't know anything)
I would say that humans do have a 'box of symbols' in real life: language. That symbolic system is normative: humans use it according to the rules that have been set down for its use.
The meaning of a word is given by an explanation of meaning, and an explanation of meaning is a rule for the use of the word explained, a standard of correct use.
That being the case, I don't understand why one would think the humans in the thought experiement couldn't use those 'symbols' to describe what they have observed since they have been provided an explanation of how they are to be used.
Thanks for providing this interesting thought experiment.
SteveK,
I will take a shot at the Searle-ish matter but, first, a couple of preliminary points.
Physicalists deny effectiveness to the conscious experience. Physicalists assert that consciousness is epiphenomenal, inert, non-contributing. David acknowledged as much when he said "I think our conscious selves are observers of our own responses" rather than effectors of anything. Accordingly, the fact that humans exhibit consciousness is supposedly a wholly irrelevant matter if - or maybe especially if - consciousness is nothing but observation. There need be no "pre-established rules" initially with regards to the symbols. In that case, the symbols are not inherently symbols; therefore, the box contains distinguishable objects. These objects can become symbols by something like the following:
Let us say that there are a chair, a table, and a rug in the room with two people who do not communicate via their voices. One person picks up an 'x' and places it on the chair, picks up a 'y' and places it on the table, and picks up a 'z' and positions it on the rug. 'x' becomes a symbol when the (brain of the) other person incorporates the association of 'x' and chair, then likewise for the other objects and their associated symbol objects. The two people can increase the number of symbols employed in something like the following manner. The second person puts the chair on the table and places the 'x' on top of the 'y' to indicate the chair is on the table. Then one of the people gets the idea to introduce another symbol ('+') to represent 'on top of'; so, 'x' '+' 'y' symbolizes chair on top of table. Then the '-' gets used to show 'y' '-' 'x' to indicate the table is under the chair, etc., etc. At some point, it may be necessary to create a new object to serve some symbolic function, but that object only becomes effective as symbol via incorporation by the other brains using the symbols/signs/representations. Here the input to the brain is by means of light waves; the supposed pattern can be further confirmed by placing the chair on top of the rug and the brain thinks it sees agreement on the part of the other person when, after that placement, the other person organizes the symbols as 'x' '+' 'z' (likewise, 'z' '+' 'x' or 'x' '-' 'z' when the rug is placed on top of the chair).
The above should be sufficient to indicate that StardustyPsyche (SP) is correct to say that, from a physicalist perspective, "'rational inference' and 'mind' are just names [symbols] applied to sets of non-rational causes."
One interesting tangent to the above regards the matter of genius - the generating of something new, such as a new symbol or even the notion of communicating via symbols. According to physicalism, the new is inevitable, nomologically necessary. The non-reductive attributing of activities to a brain is a matter of convenience, because the actual activities are a far more reductive matter. SP is on point when he says, "Victor is asserting that on materialism one inevitably must be an eliminativist in order to express materialism coherently. Victor is correct about that." A thoroughly coherent physicalist explanation requires expressions in far more reductive terms than can be (dare I say ever will be) provided. Of course, coherence is not a problem unique to physicalism.
The ultimate problem for physicalism - which is to say eliminativism - regards what is left after all the eliminating. This problem is typically avoided by resorting to what I have described as dualistic thinking. That does not have to mean substance dualism; it could mean something in line with SP's distinction between that which exists and that which is real wherein importance would reside in what SP refers to as the real.
Michael,
That's a nice illustration of how one can teach another to use a symbolic system. Of course, any explanation of meaning (or of use) such as this one can be misunderstood. When asked to give information about what is in the room the person fails to do so then we know that further instruction is necessary.
What is happening in the brain (or the mind) is really irrelevant to knowing one has learned to use such a symbolic system.
BM,
How does a representation come about and how does it do pattern recognition?
We know from physiology that brains consist of interconnected neurons. This has led to the study of neural networks simulated in digital computers, in particular their ability to 'learn' to recognise patterns. So it's not too great a stretch to think that children learn how to associate the appearance of things of various kinds with the names of the kinds by means of real neural networks in their brains. As you say, there is a 'taking in' of something external which becomes internalised in the structure and synaptic connections of the network. I don't think it unreasonable to call this 'representation'.
Representations that don't represent.
We can learn to recognise solid square objects and call them 'squares'. It's plausible that the network that recognises solid square objects will also respond in part to a square outline or four separate objects placed at the corners of a square. But we don't have a word for this. We talk perhaps about a 'square shape' acknowledging that this term does not denote a material object. We have a representation, somewhat like the impossible trident drawing, that doesn't represent a material thing. Reflecting on this we ask ourselves, What does it represent then? Squareness? Cue two millennia of debate.
...there will be a series of bits in memory that represent the common essential leftovers and one can call squareness.
I would not put it like that. We are not talking about leftovers from some process of discarding. The recogniser is something constructed in the brain by a process of exposure to many instances of squares and non-squares.
I don't understand how this somehow means that a belief rationally inferred can also be fully explained in terms of non-rational causes.
If rational inference is pattern matching and substitution over sentences, and visual recognition is an activity possible for a neural network, it's plausible that rational inference is also possible for a neural network.
David,
Even if I agreed with your last sentence the machine would not cause the machine to believe anything because machines are not the sort of things that can believe.
But leaving that aside there are more fundamental issues. Maybe you can help me understand.
If the only things that actually exist are fundamental particles then there really are no 'things' such as machines. There would be no principled reason to distinguish between the particles that the machine consists of and the those of the table is sits on or the air that surrounds it or even you. Why should I not think that half an adder, 1 cubic meter of air and 2 pounds of table are a thing rather than the parts of 3 separate things?
It seems the materialist allows himself the substance concept on the one hand to explain how materialism can account for "things" such as "brain" activity and on the other hand denies that there are really any sorts of things we call brains at all.
It also seems to me that all the talk about representation, learning, pattern matching etc all start with the assumption there are such things as substances and we are an instance of one.
The same goes for arguing for materialism from the standpoint of the study of physics. The subject requires a mind and a substantial being to even begin with. So it seems incoherent to argue that minds and substantial beings do not exist as shown by physics. If true, then no one then is really around to make the argument.
Michael,
"There need be no "pre-established rules" initially with regards to the symbols"
The rules are intended to represent physical laws, at least that's the way I understand the thought experiment. If that is the case then in my example using humans, the rules (physical laws) are creating the symbols AND they are arranging the symbols. Rational inference involves a specific ordered arrangement of facts (represented by symbols).
In my examples I didn't intend for the humans to know the relationship between the symbols and the external objects in the room. Going back to David's adding circuit, the circuit has no knowledge of numbers or addition and yet David claims it is adding. Impossible. In your example the humans know the relationship and I don't think that is possible given materialism.
Because the rules create the specific ordered arrangement of symbols the human participants aren't performing a rational inference. It's being done for them. Using your objects in the example that you gave, the rules might create the arrangement "x y - + z". Is that a rational inference involving the facts?
Much more can be said but I will stop here for now.
BM, Is anyone claiming only fundamental particles exist? Not I. Trees, houses, and men seem to me to have a much better claim on existence than quarks and electrons! I'm happy to call such things 'substances', though I prefer 'material things' or 'material objects' just to avoid the implications of 'substance' within Aristotelian theory. I am less sure about minds as things or entities. I am more inclined to say that mind is what brains do, but this is vague and contentious.
Why do we distinguish between objects and their surroundings? We know that living things are continually exchanging matter with their environment. We are like persistent eddies in the great flux. Maybe we make the distinction because our brains tell us to, as it were.
SteveK, More specifically, I say that the adder exemplifies or realises a pattern we call 'addition'. It clearly is not a mind, that is, it does not exemplify mindedness, and all that that entails.
SteveK, Searle's target with the Chinese Room argument I take to be the 'strong AI' claim that intelligent behaviour can be achieved with a stored program digital computer, or maybe that brains work 'just like' such machines. I cannot believe the latter. Such systems are extremely fragile in the face of tiny changes to their structure. Brains have evolved and further developed from babyhood and are, in contrast, very, very robust. The only way I think a SPD computer might 'do' intelligent behaviour is by simulating a physical brain. That is, by modelling the physical processes going on in a brain just as weather forecasting systems model the atmosphere.
David,
Thanks for the clarification regarding physical objects. It helps me distinguish your particular views from other typical versions of materialism.
In this case it seems you would agree that physical entities exist as something, over and above/different from, the parts they consist of in addition to the parts they consist of. What could that physical something be?
SteveK,
The rules are intended to represent physical laws ... Much more can be said but I will stop here for now.
Ah, I have a better sense of your point(s). I will take up a bit of that much more with what basically amounts to stream of consciousness.
Physicalism does not actually require laws, and physicalists do not have to insist on laws. The very notion of laws of physics really can be appreciated as a semantic convenience (even if it is one which introduces some inaccuracy). There are no laws or anything actually guiding or directing what occurs. The so-called laws are simply descriptions of what occurs regularly and predictably given particular conditions. Brief tangent: Given the block space-time notion, nothing actually occurs, and the ultimate description is all just always is. Of course, there are no descriptions any more than there are patterns if there is no awareness of occurrences. (It should be apparent that this relates to the matter of: No minds; no truths.) Patterns are generated from/by awareness so that talk of patterns is talk of awareness (and its conditions/context). If rationality including its inferences is entirely a matter of pattern recognition which abstracts away/eliminates content such that content is irrelevant to the rational pattern, then rationality and its inference are merely a matter of form (a formal matter, a formality) and are empty of anything like the awareness which generated the pattern dubbed rationality. Being able to recite a text is not identical to presenting the meaning (which is more closely associated with content rather than form/pattern). Rote memorization even with repetition is not an indication of knowledge regarding meaning/content. In and of itself, reproduction of a form/pattern (including rationality to the extent that it is merely a formal matter) is not an indicator of nor dependent upon awareness. Rationality as pure form/formality requires no understanding. However, where rationality as form(ality) is expected, then an awareness is required to do the abstracting necessary to present the awareness in the form required. Rationality as form(ality) can be mimicked far more easily than can be awareness. Under a physicalism which asserts consciousness as epiphenomenal, awareness is always other than conscious thinking.
BM, I would agree? Not entirely. For artifacts like clocks and cars the parts are objects in their own right assembled into a whole. For organic things that have grown, parts are more conceptual and are not objects in their own right. They have vague boundaries and are not cleanly separable from the whole.
David,
I understand your distinction and I agree with you in the aspect you're addressing. But the parts I was referring to where things like atoms or whatever the more fundamental material each category consists of.
OK. For the sake of discussion let's say that the parts are molecules on the understanding that there is flux at this level too, with molecules exchanging parts of themselves with others in chemical reactions. This makes the identification of molecules as persistent individuals somewhat tricky.
David,
So do I take you as holding that there is no such thing as a base "material"? You mentioned that we are eddies in the "flux". What does this flux consist of?
Your guess is as good as mine. Matter in motion.
How can I tell if it is in motion if I can't tell what it is? Some thing has to be moving from on place to another, no?
I'm asking because it seems we can directly or indirectly sense some objects but not this stuff? It's not that I disagree, I'm just curious on your take.
Seeing an object is seeing stuff. Seeing stuff in very small quantities is harder. This is not unfamiliar!
Right.
Would I be correct to say that indetectable stuff forms into things that we can detect that look and behave in various ways according to the formation of the stuff?
I would say so. Is that controversial, do you think?
Maybe.
What if I called the indetectable stuff prime matter and proceeded like this?
When prime matter is combined with forms we end up with different material objects that look and behave in particular ways. These combinations are what we call substances.
If the previous formulation is not controversial then this second one shouldn't be either. But I suspect it is because of the words I used.
Do go on.
Can I assume that the particular behaviors of a particular type of material object seem to be regular features rather than merely by chance?
Sorry if these questions seem basic but I can see I've misunderstood your positions and I'm sure that contributed to our talking past each other.
In general, yes. But what about our behaviour? That can be irregular but not by chance.
If it is irregular then how could we know it is not by chance?
Do you mean that things sometimes happen that seem to be by chance? Or something like 2 different humans may chose different actions in similar circumstances or even that the same individual may behave differently in similar circumstances at different times.
If it is the latter then we could ask the individual(s) why they did what they did. They would presumably give us reasons. Couldn't then conclude that humans regularly behave according to their reasons?
Post a Comment