Proxytype theory and/or Objectivist epistemology

[dan2100]

What are concepts? Where do they come from? What do they do? How do they work and work together? Jesse J. Prinz’s Furnishing the Mind: Concepts and Their Perceptual Basis tries to answer these questions in light of modern epistemology, cognitive science, linguistics, psychology, and neuroscience. This broad approach to these problems challenges any reviewer coming to the task with less than an encyclopedic knowledge of these fields. Yet as Prinz seems to score highly both in defending his particular theory, proxytypes (it's unfortunate he uses "proxytype" as this will likely lead to some confusion with "prototype" – and prototype theory is a major rival to his proxytype theory), and concept empiricism in general, this reviewer, lacking such wide-ranging knowledge, will still make the attempt.

The order of battle for Prinz starts with a set of criteria any viable theory of concepts would have to meet. Next, he measures other major theories against these criteria. After this, he presents his proxytype theory, evaluating in light of the criteria, prior theories, and evidence from many quarters.

Let me just open this discussion by mentioning his six criteria – or "desiderata" as he calls them. They are a checklist a theory of concepts should pass. It’s possible no theory will close the field, but missing any of the six criteria, it seems, means a theory is not even in the running.

The six are scope, intentional content, cognitive content, acquisition, categorization, compositionality, and publicity. Scope should be the easy one for the amateur philosopher: any theory of concepts should be able to handle all the various concepts posited. Of course, one way of special pleading for a theory is to redefine some alleged concepts as not the real McCoy. Prinz does do this and accepts the standard menu of concepts from "dog" to "electron" to "justice" to "disjunction." Any decent theory shouldn’t stumble on a concept like "virtue." (Later in the book, he backslides in one case. Cf. p183 where he states that "negation" might not be a concept. However, I see this as him stumbling and not his theory failing. Also, one might argue that a general theory of concepts has to meet this criterion, but a specialized one need not. While this seems a cogent reply, Prinz aims at a general theory of concepts and not one limited to, say, just concepts dealing directly with perception or with logical operations.)

For concept empiricism, scope can be especially vexing. After all, beyond obviously "empirical" concepts, such as "dog" and "yellow," many concepts seem to have little empirical content, such as the aforementioned "disjunction" and "virtue" – not to mention "infinite," "transcendent," and "ghost." Philosophers like Ayn Rand might get away with "abstractions from abstractions," but different routes to different higher-level concepts make this harder to defend. (See Rand's ITOE 2/e, pp19-28.)

Intentional content is a bit more technical. It is the notion that concept refers to something outside itself. In other words, the concept "dinosaur" refers to dinosaurs. (As one might guess, any form of concept empiricism will have at least some concepts that refer directly to experience, though wouldn't a non-empirical theory of concepts, e.g., Fodor's, do likewise?)

Reference, though technical, seems easy compared to his criterion of cognitive content. Prinz relies on Gottlob Frege’s distinction between the sense and reference of a term. Simply put, Frege splits meaning into two components. One is what a term refers to – the thing picks out in the world. This is its reference. The other is what a term means as distinguished from other terms – for Prinz, what a concept’s content is that distinguishes it from other concepts.

The archetypal example is having two terms for Venus – the Morning Star and the Evening Star. In Prinz’s view, this makes for differing cognitive content – different senses. Even though both terms mean Venus, they mean Venus in slightly different ways. Needless to say, for him, any theory of concepts would have to deal with this sort of cognitive content. (Here I must admit to being unconvinced by Prinz’s examples. I'm trying to work out my ideas on Frege's sense and reference, but that's a post for another time.)

A decent theory must "ultimately support a plausible explanation of how concepts are acquired." (p8) Prinz thinks there are two parts to this criterion: one ontogenetic, the other phylogenic – one about how individual concepts are acquired, the other about how the conceptual faculty itself evolved.

Some might stop with the ontogenetic side of the story – and, indeed, this seems to be where Rand stops. (In my view, Rand does not really present a phylogenic explanation of concepts, though she does offer that the conceptual faculty is humanity’s basic means of survival. This might seem to put her in the Darwinian camp, but she resists letting Darwinism intrude into philosophy.) One can heed the warning that wedding an epistemological theory too closely with specific evolutionary theories or findings could end in a faddish or simplistic explanation. As a purely practical matter, other might conflate this element of the theory with its essence and reject the whole theory if this element proves wrong. However, a theory of concepts should be able to either fit into the latest science or explain why it’s outside such science. (See Lionel Robbins' 1932 book An Essay on the Nature and Significance of Economic Science, chapter 5. Robbins argues that economics, like mathematics, should be above the particulars of empirical findings. See also Paul Benacerraf's 1965 paper "What numbers could not be" for why a set theoretic reduction of numbers doesn’t seem to actually reduce numbers to sets. In the same way, it appears other fields might be immune to such reductions. Whether this is a fatal blow to reductionism in general remains to be seen.)

This aside, the ontogenetic part of the criterion seem uncontroversial: how an individual comes to acquire a concept needs to be explained.

For Prinz, categorization encompasses "mechanisms for forming beliefs about what things fall under our concepts" and is, obviously, the "epistemic counterpart" to reference. (p9) Like acquisition, this has two different aspects: category identification and category production. The first refers to identifying the category to which something belongs – as in the choice: "True or false: canaries are birds." (p9) (This maps onto "extension" as usually understood. Simply put, a concept’s extension is "made up of all those entities… which fall under the concept." Flew A Dictionary of Philosophy, p117.)

The second is "intension" – which properties make up a category, such as "having feathers" for birds. In other words, while identification is about whether something belongs to a category – whether a canary is a bird – production is about why it belongs – the things that make birds birds.

Here Prinz mentions typicality – that more typical members of a category are easier to categorize. He also observes that mid-level concepts are "privileged" in many instances. He uses the example of classifying an object as a "dog" – rather than a "rottweiler" or a "mammal." (p10) (It seems this is the case because of context and unit economy – two key concepts from Rand. Most people would acquire “dog” earlier and use it more often because this usually and quickly identifies the object in everyday life. A dog will likely be among other pets or things, but not always other dogs. Thus, it’s economical to point to the neighbor’s dog – rather than the neighbor’s pet or the neighbor’s animal, or the neighbor’s rottweiler – in most circumstances. See also pp163-4.)

Compositionality means the ability to combine concepts together. Hume’s "golden mountain" is the classic example. (Hume An Inquiry Concerning Human Understanding. Prinz does not bring up the golden mountain example at this point, but does in the next chapter when discussing Hume’s "imagism.") A good theory of concepts should account for this capability – indeed for its "hyperfertility" as Prinz aptly puts it.

Publicity is the ability for different people to share the same concepts. (Prinz also adds the ability for one person to hold the same concept across time – something he later elaborates on when discussing memory.) There’s room for much debate here, though his point seems valid: any decent theory would have to explain this feature of concepts. (A point of contention is deciding when two people actually share the same concept. If your concept of "fish" includes whales while mine doesn’t, are we really sharing the concept of fish? The same applies to deciding when one person actually has the same concept at different times. If you held a concept of "fish" that included whales yesterday, but today you’ve been enlightened so as to exclude whales, do you really have the same concept on both days? These examples might seem to be purely nit picking, but until two people do extensive note trading, how can they be sure that any of their concepts are really shared? The same goes for the single individual: How can she be sure any of her concepts are the same from one day to the next?)

Prinz considers sharing along conceptual and intentional lines as well – and the use of sharing in explaining linguistic communication and behavior.

I believe that's enough to chew on for now -- rather than extend this out to be a full length review. Comments?

[Guyau]

Daniel Ust, welcome to Objectivist Living!

~~~~~~~~~~~~~~~~

Re: #3

Yes, makes sense. Does Prof. Prinz connect logical connectives like disjunction to motor exerience?

I have not studied this book yet. Your review is informative and a good invitation. I appreciate your initiating assessment of Rand’s conception of concepts against the desiderata of Prinz. I see there is a review by James Beebe complementary to yours.

Edited March 3, 2010 by Stephen Boydstun

[dan2100]

Daniel Ust, welcome to Objectivist Living!

Thanks Stephen!

I hope my partial review makes sense.

[Selene]

Dan:

Welcome to OL.

I gather Prinz's "new empiricism" is controversial.

"This article examines the role of concepts in the so-called 'new' empiricism that is currently emerging from the writings of Gilles Deleuze. It asks what concepts are, and how they might be put to work to present the 'pure difference' of the empirical world. In addressing these questions, a number of parallels and contrasts are drawn between the writings of Deleuze and Max Weber. It is shown that many of Deleuze's key arguments about concepts- in particular, that they are pedagogical, multiple, networked and problem-oriented in basis - are anticipated by Weber's sociological methodology of concept formation. This leads, finally, to a consideration of whether the creation of concepts as a practice belongs primarily within the domain of philosophy (as argued by Deleuze), or if it is a key part of social scientific work more generally."[ Nicholas Gane UNIVERSITY OF YORK, UK, from the of European Journal of Social Theory]

Adam

[dan2100]

Dan:

Welcome to OL.

I gather Prinz's "new empiricism" is controversial.

"This article examines the role of concepts in the so-called 'new' empiricism that is currently emerging from the writings of Gilles Deleuze. It asks what concepts are, and how they might be put to work to present the 'pure difference' of the empirical world. In addressing these questions, a number of parallels and contrasts are drawn between the writings of Deleuze and Max Weber. It is shown that many of Deleuze's key arguments about concepts- in particular, that they are pedagogical, multiple, networked and problem-oriented in basis - are anticipated by Weber's sociological methodology of concept formation. This leads, finally, to a consideration of whether the creation of concepts as a practice belongs primarily within the domain of philosophy (as argued by Deleuze), or if it is a key part of social scientific work more generally."[ Nicholas Gane UNIVERSITY OF YORK, UK, from the of European Journal of Social Theory]

Adam

Thank you for your welcome.

Regarding it being controversial, yes, I think it is, though Prinz does not appear to be influenced by the works or views of Deleuze. Instead, he seems to be harking back to the British empiricists, though updating their views in light of modern philosophy, linguistics, and science as well as the intervening period of criticism of those earlier views. He's not merely warming over, say, Locke and Hume. (I'm not a Deleuze scholar, but it seems to me like his epistemology is merely an afterthought to his ideology. Perhaps someone might correct me on this.)

[Selene]

Dan:

Thanks.

This is a discipline that I have always been intrigued by, but am not familiar with current iterations, until your post.

What makes his approach controversial in "academia"?

I have always been amazed by what is controversial in "academia" particularly when I was part of it. I was considered quite rude, but I did not lose many serious debates within the "community."

Adam

[dan2100]

Dan:

Thanks.

This is a discipline that I have always been intrigued by, but am not familiar with current iterations, until your post.

What makes his approach controversial in "academia"?

I have always been amazed by what is controversial in "academia" particularly when I was part of it. I was considered quite rude, but I did not lose many serious debates within the "community."

Adam

My understanding is conceptual empiricism -- under which proxytypes falls -- is a bit controversial in the field. I don't attend their conferences, so I'm only guessing this based on what I've read. Since some rivals to Prinz's theory can be empiricist, I don't think empiricism is completely out of fashion at this time. (Prinz's definition of concept empiricism is the view that "concepts are copies or combinations of perceptual representations." I don't think that would be controversial for Objectivists -- save for the use of "perceptual representations" which might make them think of representationalism a la Descartes.)

The rivals I mean are things like prototype theory and exemplar theory. In prototype theory, concepts are formed from typical feature of members that fall under a concept. Think of a typical features of a dog -- such as having fur, barking, wagging its tail -- to build one's concept of "dog." You might gather where there's a problem with this view: what's typical to you might not be typical to me, yet it's likely both us use "dog" to mean the same thing.

Exemplar theory is similar to prototype theory, but uses one or more examples of something to form the concept -- sort of like using your neighbor's dog to build your concept of dog. In this case, one might say, the examplars are literally stored in memory to use as the concept. Anyhow, rather than go into further detail, I think you can see these are both empiricist in outlook -- or that they mesh well with conceptual empiricism. And both these theories have their respective followings.

Despite my starting this discussion off, I haven't mention what proxytypes are and how they differ from other views -- even the two I just mentioned. Let me save that for another post.

[BaalChatzaf]

What are concepts? Where do they come from? What do they do? How do they work and work together? Jesse J. Prinz's Furnishing the Mind: Concepts and Their Perceptual Basis tries to answer these questions in light of modern epistemology, cognitive science, linguistics, psychology, and neuroscience. This broad approach to these problems challenges any reviewer coming to the task with less than an encyclopedic knowledge of these fields. Yet as Prinz seems to score highly both in defending his particular theory, proxytypes (it's unfortunate he uses "proxytype" as this will likely lead to some confusion with "prototype" – and prototype theory is a major rival to his proxytype theory), and concept empiricism in general, this reviewer, lacking such wide-ranging knowledge, will still make the attempt.

The order of battle for Prinz starts with a set of criteria any viable theory of concepts would have to meet. Next, he measures other major theories against these criteria. After this, he presents his proxytype theory, evaluating in light of the criteria, prior theories, and evidence from many quarters.

Let me just open this discussion by mentioning his six criteria – or "desiderata" as he calls them. They are a checklist a theory of concepts should pass. It's possible no theory will close the field, but missing any of the six criteria, it seems, means a theory is not even in the running.

The six are scope, intentional content, cognitive content, acquisition, categorization, compositionality, and publicity. Scope should be the easy one for the amateur philosopher: any theory of concepts should be able to handle all the various concepts posited. Of course, one way of special pleading for a theory is to redefine some alleged concepts as not the real McCoy. Prinz does do this and accepts the standard menu of concepts from "dog" to "electron" to "justice" to "disjunction." Any decent theory shouldn't stumble on a concept like "virtue." (Later in the book, he backslides in one case. Cf. p183 where he states that "negation" might not be a concept. However, I see this as him stumbling and not his theory failing. Also, one might argue that a general theory of concepts has to meet this criterion, but a specialized one need not. While this seems a cogent reply, Prinz aims at a general theory of concepts and not one limited to, say, just concepts dealing directly with perception or with logical operations.)

For concept empiricism, scope can be especially vexing. After all, beyond obviously "empirical" concepts, such as "dog" and "yellow," many concepts seem to have little empirical content, such as the aforementioned "disjunction" and "virtue" – not to mention "infinite," "transcendent," and "ghost." Philosophers like Ayn Rand might get away with "abstractions from abstractions," but different routes to different higher-level concepts make this harder to defend. (See Rand's ITOE 2/e, pp19-28.)

Intentional content is a bit more technical. It is the notion that concept refers to something outside itself. In other words, the concept "dinosaur" refers to dinosaurs. (As one might guess, any form of concept empiricism will have at least some concepts that refer directly to experience, though wouldn't a non-empirical theory of concepts, e.g., Fodor's, do likewise?)

Reference, though technical, seems easy compared to his criterion of cognitive content. Prinz relies on Gottlob Frege's distinction between the sense and reference of a term. Simply put, Frege splits meaning into two components. One is what a term refers to – the thing picks out in the world. This is its reference. The other is what a term means as distinguished from other terms – for Prinz, what a concept's content is that distinguishes it from other concepts.

The archetypal example is having two terms for Venus – the Morning Star and the Evening Star. In Prinz's view, this makes for differing cognitive content – different senses. Even though both terms mean Venus, they mean Venus in slightly different ways. Needless to say, for him, any theory of concepts would have to deal with this sort of cognitive content. (Here I must admit to being unconvinced by Prinz's examples. I'm trying to work out my ideas on Frege's sense and reference, but that's a post for another time.)

A decent theory must "ultimately support a plausible explanation of how concepts are acquired." (p8) Prinz thinks there are two parts to this criterion: one ontogenetic, the other phylogenic – one about how individual concepts are acquired, the other about how the conceptual faculty itself evolved.

Some might stop with the ontogenetic side of the story – and, indeed, this seems to be where Rand stops. (In my view, Rand does not really present a phylogenic explanation of concepts, though she does offer that the conceptual faculty is humanity's basic means of survival. This might seem to put her in the Darwinian camp, but she resists letting Darwinism intrude into philosophy.) One can heed the warning that wedding an epistemological theory too closely with specific evolutionary theories or findings could end in a faddish or simplistic explanation. As a purely practical matter, other might conflate this element of the theory with its essence and reject the whole theory if this element proves wrong. However, a theory of concepts should be able to either fit into the latest science or explain why it's outside such science. (See Lionel Robbins' 1932 book An Essay on the Nature and Significance of Economic Science, chapter 5. Robbins argues that economics, like mathematics, should be above the particulars of empirical findings. See also Paul Benacerraf's 1965 paper "What numbers could not be" for why a set theoretic reduction of numbers doesn't seem to actually reduce numbers to sets. In the same way, it appears other fields might be immune to such reductions. Whether this is a fatal blow to reductionism in general remains to be seen.)

This aside, the ontogenetic part of the criterion seem uncontroversial: how an individual comes to acquire a concept needs to be explained.

For Prinz, categorization encompasses "mechanisms for forming beliefs about what things fall under our concepts" and is, obviously, the "epistemic counterpart" to reference. (p9) Like acquisition, this has two different aspects: category identification and category production. The first refers to identifying the category to which something belongs – as in the choice: "True or false: canaries are birds." (p9) (This maps onto "extension" as usually understood. Simply put, a concept's extension is "made up of all those entities… which fall under the concept." Flew A Dictionary of Philosophy, p117.)

The second is "intension" – which properties make up a category, such as "having feathers" for birds. In other words, while identification is about whether something belongs to a category – whether a canary is a bird – production is about why it belongs – the things that make birds birds.

Here Prinz mentions typicality – that more typical members of a category are easier to categorize. He also observes that mid-level concepts are "privileged" in many instances. He uses the example of classifying an object as a "dog" – rather than a "rottweiler" or a "mammal." (p10) (It seems this is the case because of context and unit economy – two key concepts from Rand. Most people would acquire "dog" earlier and use it more often because this usually and quickly identifies the object in everyday life. A dog will likely be among other pets or things, but not always other dogs. Thus, it's economical to point to the neighbor's dog – rather than the neighbor's pet or the neighbor's animal, or the neighbor's rottweiler – in most circumstances. See also pp163-4.)

Compositionality means the ability to combine concepts together. Hume's "golden mountain" is the classic example. (Hume An Inquiry Concerning Human Understanding. Prinz does not bring up the golden mountain example at this point, but does in the next chapter when discussing Hume's "imagism.") A good theory of concepts should account for this capability – indeed for its "hyperfertility" as Prinz aptly puts it.

Publicity is the ability for different people to share the same concepts. (Prinz also adds the ability for one person to hold the same concept across time – something he later elaborates on when discussing memory.) There's room for much debate here, though his point seems valid: any decent theory would have to explain this feature of concepts. (A point of contention is deciding when two people actually share the same concept. If your concept of "fish" includes whales while mine doesn't, are we really sharing the concept of fish? The same applies to deciding when one person actually has the same concept at different times. If you held a concept of "fish" that included whales yesterday, but today you've been enlightened so as to exclude whales, do you really have the same concept on both days? These examples might seem to be purely nit picking, but until two people do extensive note trading, how can they be sure that any of their concepts are really shared? The same goes for the single individual: How can she be sure any of her concepts are the same from one day to the next?)

Prinz considers sharing along conceptual and intentional lines as well – and the use of sharing in explaining linguistic communication and behavior.

I believe that's enough to chew on for now -- rather than extend this out to be a full length review. Comments?

A first rate essay! You do know your stuff. Just one little nit. Natural integers can be reduced to sets. Let S be a set. Then the cardinal of S is the class of sets S* which can be put in one to one correspondence with S. Integers are cardinals of finite sets.

Please share more of your thoughts. You really know what you are talking about.

Ba'al Chatzaf

[dan2100]

A first rate essay! You do know your stuff. Just one little nit. Natural integers can be reduced to sets. Let S be a set. Then the cardinal of S is the class of sets S* which can be put in one to one correspondence with S. Integers are cardinals of finite sets.

Please share more of your thoughts. You really know what you are talking about.

Ba'al Chatzaf

Thank you, though this is not so much an essay, so much as maybe the beginning of one -- of a review essay. I don't know how much further I'll get with this here. Also, I'm merely summarizing and commenting on Prinz's book -- not doing much intellectual heavy-lifting here.

Regarding your "little nit," have you read Benacerraf's "What numbers could not be"? It seems to call into question set theoretic reductions -- despite the fact that they otherwise seem to work, are quite useful, and seem intuitively appealing.

Regards,

Dan

[Selene]

Yes Ba'al:

I like his approach. Good mind.

Adam

[dan2100]

Yes Ba'al:

I like his approach. Good mind.

Adam

Thank you, but don't you fear this praise might go to my head?

[Selene]

Yes Ba'al:

I like his approach. Good mind.

Adam

Thank you, but don't you fear this praise might go to my head?

Nope, I can make a head shot. Learned to shoot when I was real young!

LOL

Sniper kitten

Edited March 2, 2010 by Selene

[dan2100]

Daniel Ust, welcome to Objectivist Living!

~~~~~~~~~~~~~~~~

Re: #3

Yes, makes sense. Does Prof. Prinz connect logical connectives like disjunction to motor exerience?

I have not studied this book yet. Your review is informative and a good invitation. I appreciate your initiating assessment of Rand's conception of concepts against the desiderata of Prinz. I see there is a review by James Beebe complementary to yours.

I have to go back to the book as I forget at this point. I started writing a review of it in 2006 and didn't complete that. What I've placed here is just culled from that.

I also recommend Prinz's book on emotions, but that's a topic for another discussion -- one I hope we can have here soon.

~~~~~~~~~~~~~~~~

I'm in midst of moving, so I had to retrieve my copy of Furnishing the Mind from my new library to more fully respond. Prinz does address the issue of logical connectives in chapter 7 "The Perceptual Basis." There he examines counterexamples, but it's more of a list of strategies for dealing with them -- concepts of logic (disjunction, etc.), mathematics, and other seemingly non-empirical ones -- than anything else. And, re-reading it very quickly this morning, he doesn't list a motor experience strategy. (A few years ago, I emailed Prinz regarding mathematical concepts -- specifically citing Kitcher's The Nature of Mathematical Knowledge and your and my review of that work, but I didn't get a substantive reply and he was unfamiliar with Kitcher's work.) He doesn't commit to any strategy in the end, but in this chapter is merely out to show that the counterexamples can be countered by some strategy rather than to settle on a particular one.

I actually had some notes on his not using processes for his proxytype theory. I want to raise some of my criticisms later on this site, but three I think should be raised now is he doesn't deal explicitly with "abstractions from abstractions," he doesn't have perceptions of processes exactly down (which would give him the whole Kitcherian shebang, no?), and, at one point, he even goes so far to offer that "negation" might not be a concept (p183). This last seems like limiting scope -- one of his criteria -- to suit the theory.

Edited March 4, 2010 by Dan Ust

[dan2100]

I'm in midst of moving, so I had to retrieve my copy of Furnishing the Mind from my new library to more fully respond. Prinz does address the issue of logical connectives in chapter 7 "The Perceptual Basis." There he examines counterexamples, but it's more of a list of strategies for dealing with them -- concepts of logic (disjunction, etc.), mathematics, and other seemingly non-empirical ones -- than anything else. And, re-reading it very quickly this morning, he doesn't list a motor experience strategy. (A few years ago, I emailed Prinz regarding mathematical concepts -- specifically citing Kitcher's The Nature of Mathematical Knowledge and your and my review of that work, but I didn't get a substantive reply and he was unfamiliar with Kitcher's work.) He doesn't commit to any strategy in the end, but in this chapter is merely out to show that the counterexamples can be countered by some strategy rather than to settle on a particular one.

I actually had some notes on his not using processes for his proxytype theory. I want to raise some of my criticisms later on this site, but three I think should be raised now is he doesn't deal explicitly with "abstractions from abstractions," he doesn't have perceptions of processes exactly down (which would give him the whole Kitcherian shebang, no?), and, at one point, he even goes so far to offer that "negation" might not be a concept (p183). This last seems like limiting scope -- one of his criteria -- to suit the theory.

Stephen's and my review of Kitcher's book is online at:

http://objectivity-a...number6.html#55

Edited March 4, 2010 by Dan Ust

[Guyau]

.

You were a bright spot here, Dan.

Sorry it did not work out.

.

[Michael Stuart Kelly]

Stephen,

For the record, Dan can come and go as he pleases.

There is no restriction against him. Never was.

There might have been had he pursued the threat he made against me in his last post, but I interpreted that as anger-talk or posturing. That attitude would have had to go a lot further before I would have considered a restriction since it was atypical.

Michael