Anti-Abstractionism and Philosophy of Language
(I) Locke on Abstraction
-According to Locke, simple ideas come in through 'sensation' and 'reflection' (these are the two basic sources of our ideas).
-The mind can then go to work on these ideas. There are several procedures: compounding and enlarging, comparing, and abstraction.
-For Locke, abstraction involves moving from the particular to the general. For example, one may have complex ideas of Peter, James, and Mary. One may then move from these particular ideas (and various other ideas like it) to an abstract idea of human beings in general.
-Locke suggests that this process of abstraction involves 'leaving out' the differences and retaining the similarities. So, for example, in moving from particular ideas of Peter, James, and Mary to the (abstract) idea of human being in general, one 'leaves out' all of the features idiosyncratic to Peter, James, and Mary, and retains only those features which are common to all.
-Abstraction plays an important role in Locke's philosophy of language. He claims that general terms (such as 'human being') get their meaning by having an abstract idea attached to them.
-According to Locke, the chief difference between animals (brutes) and human beings is that the latter are capable of abstraction while the former are not.
(II) Contrasting Interpretations of Berkeley's Rejection of Abstraction
(1) The Imagist Reading. Descartes distinguishes between ideas of the 'pure intellect' and ideas produced by the imagination. For example, we cannot conjure up the visual image of a chiliagon, but we nonetheless understand what a chiliagon is. Descartes' thought is that we have a 'pure idea' of a chiliagon. Now on this particular reading, Berkeley assumes that abstract ideas must be produced by the imagination (i.e. he ignores 'pure ideas'). He then claims that every imagined idea that we produce is of some particular object (rather than anything general) and goes on to conclude that abstraction (generalization) is impossible. Notably, if this interpretation of Berkeley is correct, then Berkeley simply begs the question because he excludes 'pure ideas' without any argument! And Locke himself would not have supposed that abstract ideas are imagined ones, either. Consequently, Berkeley would also be begging the question against Locke.
(2) Berkeley's Bad Blunder. Locke, at one point in the Essay, appears to suggest that in forming an abstract idea (of human being in general, for example) that we both (i) leave out all of the differences between the particulars; and (ii) throw in all the differences. He appears to suggest that in order to form an abstract idea (of a human being in general, for example) that we must form the idea of something that (i) is six feet tall, and five feet tall, and four feet tall, etc; and (ii) is NOT six feet tall, and is NOT five feet tall, etc. (See Locke's Essay Book 4, chapter 7, sec. 9). Now this is clearly an absurd position.
Berkeley makes fun of this passage at PHK Intro sec. 13. However, this is NOT Locke's actual position. His actual position is simply that we retain the similarities and leave out the differences. Now on this interpretation, Berkeley is supposed to misunderstand Locke by attributing to him the insane position described above. Notably, this reading would make Berkeley look fairly stupid. Notice, however, that Berkeley correctly describes Locke's actual account of abstraction in sections 8 and 9 of PHK Intro. So I doubt that this reading can be correct.
(3) Selective Attention versus Separation. When Locke says that we retain the similarities and 'leave out' the differences in abstracting from particular to general, what does he have in mind? There are two possibilities.
First, Locke might mean that we break apart the complex ideas of Peter, James, etc. and retain only the similar ideas. This is a kind of separation (just as one might imagine an arm without the rest of the human body).
Second -Locke might mean that we begin with the idea of some particular human being and then only 'focus on' the similar features while 'ignoring' the idiosyncratic ones.
One way to understand the difference is that in the first case (but not the second) it seems like a new idea is produced. Another way to understand the difference is that in the first case (but not the second) the abstracted feature is being conceptualized as existing on its own.
It is subject to debate which of these Locke means (separation or selection). However, one interpretation of Berkeley says that while Locke believes in abstraction as selection, Berkeley misreads him to mean abstraction as separation. Berkeley, in this interpretation, critiques and rejects abstraction as separation but himself accepts and believes in abstraction as selection.
This is Winkler’s Interpretation. Winkler thinks that Berkeley uses the Impossibility Argument against Locke:
The Impossibility Argument. If it is impossible in reality, it is impossible in
(A) Content Assumption: The content of a thought is determined by its object (i.e.
there is nothing more to a thought than the idea)
(B) Reality Assumption: If something is impossible, then it cannot be an object of
Thought (i.e. one cannot form an idea of it)
According to Winkler, Berkeley thinks that Locke accepts the Content Assumption and he thinks that Berkeley rejects it. One way to point the point: If the Content Assumption is true, then Mental Selective Attention is impossible. For if it were possible to attend to discrete aspects of an idea, then there would be more to a thought than its object. According to Winkler, Berkeley accepts mental selective attention but does not think that Locke accepts it.
In my own view (as I shall argue) Berkeley REJECTS CA and knows full well that Locke accepts it.
(4) Talia’s Proposed Reading of Berkeley's Rejection of Abstraction
I am unhappy with the preceding readings of Berkeley. All of them require viewing Berkeley as some one who either misunderstood Locke or who offered an exceptionally poor argument for his position. I believe that both Locke and Berkeley are committed to selective attention. However I also think there are important differences in the kind of selection. In order to see this, I want to think about the differences between Locke and Berkeley with respect to philosophy of language.
Differences Between Locke and Berkeley with respect to Philosophy of Language
(A) Locke thinks that names need to have one settled, determinate meaning (i.e. idea). A general term has an abstract idea for its meaning. According to Berkeley, names can indifferently denote a range of particular ideas. Thus a term like ‘triangle’ indifferently denotes a range of ideas of particular triangles.
(B) Locke claims that the transfer of ideas is the chief function of language. This involves a speaker using words to convey his/her ideas to an auditor. By contrast, Berkeley recognizes other important functions of language (see PHK Intro Sec. 20). Words can be used to get an auditor in a particular disposition or frame of mind, to have a particular passion, or used to incite to action. Terms can also be used as calculative devices (as in algebra). In such cases, even though the value of the term is fixed in advance, there is no need for the speaker to constantly have an idea in mind (nor is there any need for an auditor to have any idea suggested to mind).
(C) Locke thinks that the speaker always needs to have ideas 'before her mind' in order to communicate intelligibly. By contrast, Berkeley thinks that when we are reasoning in algebra (for example) we may simply use the words as calculative devices without always having the ideas 'before us'. (See PHK Intro Sec. 19). Indeed, he thinks that often general terms (such as ‘dog’) are used without specific ideas annexed to them. Thus, when I say “Helena is a dog”, I have only ONE idea (namely an idea of Helena). There is no further idea of ‘dogness’.
Mental Selective Attention \'s. Discursive Selective Attention
While Locke believes in mental selective attention, Berkeley denies that this is possible. In my reading of Berkeley, this denial is at the core of his rejection of abstractionism. What Berkeley means to reject is the very possibility of mental selective attention. Berkeley himself believes in only discursive selective attention.
What is the difference?
Recall that according to Locke, general terms get their meanings by having abstract ideas attached to them. This means that the process of abstraction (i.e. 'selective attention') is prior to language. And this suggests that we should be able to abstract without using language at all. Because of this, Locke talks about “partial ideas” or “partial considerations” which are presumably the consequence of mentally attending to only some aspects of an idea.
By contrast, Berkeley thinks that we selectively attend to some features, while ignoring others simply by mentioning some features and failing to mention others. On Berkeley's view, selective attention is NOT prior to language. On the contrary, selective attention has to do with the use of language. This is what I call 'discursive selective attention'.
(III) Berkeley's Arguments
(1) Berkeley's main argument appears to run as follows: "I try and I try and I find that I just can't form an abstract idea. Here. You try it. I bet that you can't form one either! I guess there's no such thing as abstract ideas!" On the face of it, this 'argument' seems exceptionally weak. [I can imagine Locke replying "Well, I find that I CAN form abstract ideas. So perhaps there's something wrong with your mental abilities!"]
The 'argument' seems stronger, however, when we clarify Berkeley's challenge. What we must do is to mentally selectively attend to some features while ignoring others without using language at all. So start with an idea of Helena. Now -without using any language -selectively attend to those features which all dogs have in common and leave out the differences. Can you do it?
(2) Berkeley suggests that his theory of language can explain all of the relevant phenomena that Locke's theory is supposed to explain. He also suggests that Locke's theory ends up being counter-intuitive in certain ways (whereas his own does not). So the overall verdict, according to Berkeley, should be that his own theory is superior to Locke's. One of the problems with Locke's theory, according to Berkeley, is that Locke's mental selective attention is supposed to be very difficult. Yet it is also supposed to be necessary for communication (at least vis a vis general terms). But if it is so difficult to form abstrnct ideas, then why is it that children seem to have no serious difficulties learning language?
(IV) Berkeley's Anti-Abstractionism And His Overall Project
Why does Berkeley care so much about abstract ideas? Why does he think that it is so important to argue against them?
First, Berkeley's anti-abstractionist doctrine plays a direct role in helping him achieve his overall project. Recall that Berkeley is interested in showing that much philosophical speculation involves pointless discussions about nothing. By showing that there are no abstract ideas, Berkeley hopes to rule out the following sorts of philosophical questions: "What is the Good?”, "What is virtue, really?" "What is time?” etc. According to Berkeley such questions are 'merely verbal'. See Berkeley's remarks PHK I 97-134 for a full account of how his rejection of abstraction influences his position on arithmetic, natural science, and philosophy.
The second reason that Berkeley's anti-abstractionist doctrine is so important is that it plays a key role in his argument (or at least some of his arguments) in favor of the core position. Two examples:
(1) Berkeley wants to argue that with respect to sensible things (such as colors, odors, sounds, etc.) 'to be is to be perceived'.
(2) Berkeley wants to argue that what is true of secondary qualities such as colors, sounds, and so forth is also true of primary qualities such as extension. Since color can't exist unperceived, extension can't either (Berkeley wants to argue). Central to this claim is his contention that we cannot conceive of (i.e. form the idea of) extension without secondary qualities. And the reason that we cannot do this is precisely because we cannot form an abstract idea of extension.