Re: [quickphilosophy] Fodor on Concepts IV: Circularity + Peacocke

  • From: wittrsl@xxxxxxxxxxxxx
  • To: quickphilosophy@xxxxxxxxxxxxxxx
  • Date: Mon, 4 Oct 2010 21:04:21 -0700 (PDT)

Hello once more QPers:
It seems by the end of the 1st paragraph beginning on p. 44 of his article that 
Fodor has absolved Peacocke of the crime of circular reasoning. But, then Fodor 
reflects on the situation again and renews his argument that Peacocke argues in 
a circle when he gives the rules of inference sufficient for the CONJUNCTION 
concept in schema (I) (below, and on p. 41 of Fodor's article.
Fodor's renewed argument goes like this:
1. Consider the sentence 'if John swims and Mary swims, then John and Mary swim'
2. This expresses the kind of thought the subject must find "primitively 
compelling" (in Peacocke's words) in order to possess the concept of 
3. English speakers can only be primitively compelled by this because they know 
that the word 'and' in the antecedent of (1) expresses conjunction.
4. It's only because English speakers know that 'and' expresses conjunction 
that they are compelled by arguments you get by substituting 'and' for 'C' in 
schema (I).
5. Thus, Peacocke's possession criterion requires that the subject already 
possess some other representation that expresses CONJUNCTION.
6. But, then the subject already possesses CONJUNCTION as a concept.
7. So, Peacocke is arguing in a circle.
There are a number of problems with Fodor's argument here:
(a) The sentence in (1) is not of the form of Peacocke's schema (I), which lays 
out the inference pattern required for CONJUNCTION without using 'and' at all.
(b) The inference that Fodor says the subject must make in fact is not the 
conjunction of propositions, but another proposition involving Noun Phrase 
conjunction, which is not of the form that Peacocke specifies, and which only a 
speaker of the natural language would be able to enunciate. Fodor seems to 
recognize that his argument is a strawman and has a little footnote at the 
bottom of p. 44 that says that he's only trying to "ease the exposition", but 
it's a grand leap to get from Peacocke's schema (I) to NP conjuction-containing 
sentences derived from a pair of conjoined propositions.
(c) Indeed, Fodor's schema in general does not work, because its proper 
application depends upon a knowledge of English semantics, which Peacocke's 
schema (I) does not. For example, if I take "John is a single person and Mary 
is a single person" and then by NP conjunction derive "John and Mary is a 
single person", then, well, that's wrong. 
(d) Fodor again argues by a single example in order to prove a general point. 
Plato does this in order to present Socrates making a bad argument and draw the 
reader into the thinking process. But that's a different form of philosophical 
(e) Point (4) above is also false. English speakers could be compelled not by 
the form of Fodor's false version of Peacocke's schema (I), because they could 
be compelled by the semantics and because they may have learned this 
restatement of the antecedent by rote.
(f) Point (5) does not follow. Peacocke wants to show that a subject possesses 
a concept of CONJUNCTION. But Fodor is claiming that the subject possesses 
another concept, one other than the one that the primitive compulsion is 
revealing. This appears to be a terrible blunder on Fodor's part.
(g) Well, maybe you see my point.
I've arrived at the last paragraph beginning on p. 44. Phew.

--- On Fri, 10/1/10, Ron Allen <wavelets@xxxxxxxxxxx> wrote:

From: Ron Allen <wavelets@xxxxxxxxxxx>
Subject: Re: [quickphilosophy] Fodor on Concepts IV: Circularity + Peacocke
To: quickphilosophy@xxxxxxxxxxxxxxx
Date: Friday, October 1, 2010, 11:22 AM



      As a pole around which to argue about types of circular reasoning, Fodor 
introduces another rule (R) in his polemic against Peacocke:
(R) the inference to 'p and q' is valid iff p and q are both true
Now, this appears to be circular, because the elucidation of what 'p and q' 
should mean relies on an instance of "and". Fodor goes on to say that there's 
basically (almost) nothing wrong with this, because the second instance of 
'and' is a metalanguage element. In fact, Fodor says that as long as this is a 
theory of the content of AND, it's all right to rely on an instance of AND in 
the metalanguage. But, to be precise, Fodor argues, it's not OK for a theory 
about how AND is learned to presuppose a language with the concept of 
conjuction already present. This would be a circular argument.
I'm at the middle of p. 43 right now. Gotta go get some Thai food for lunch. 
Back in a bit.

--- On Fri, 10/1/10, Ron Allen <wavelets@xxxxxxxxxxx> wrote:

From: Ron Allen <wavelets@xxxxxxxxxxx>
Subject: Re: [quickphilosophy] Fodor on Concepts IV: Circularity + Peacocke
To: quickphilosophy@xxxxxxxxxxxxxxx
Date: Friday, October 1, 2010, 9:49 AM


Hi quick group:
I'm still digging my way through Section 3.3 of Fodor's article 'Having 
Concepts'. In the last paragraph beginning on p. 41, Fodor claims that Peacocke 
(A Study of Concepts, Cambridge, MA: MIT Press, 1996) commits an equivocation 
in formulating the concept possession condition for the implicit definition 
Peacocke says that someone has to follow the CONJUNCTION scheme (below, in 
point 2. of my earlier post; it's on p. 41 of Fodor's article, where it's 
labeled (I)) by virtue of its form alone. So Fodor says, well, does that mean 
that the form is causally sufficient for the subject of Peacocke's experiment 
to find them compelling? or, if not, then is that the reason that the subject 
has for being compelled?
It seems to me, though, that this is a false dichotomy. If the form isn't 
causally sufficient, then something else is required for the cause; it doesn't 
mean that the form of the schema is the reason for the subject's compulsion. 
All Peacocke wants out of this is to say that, whatever the motivation or 
impetus for the subject to assent to the schema, it's the forms themselves, and 
not some external factor, like whether the subject agrees with or approves of 
the propositions p and q in the claim pCq, that result in the subject's 
demonstrating facility with the schema. The subject doesn't have to have a 
reason; she only has to follow the pattern of the implicit definition. And, the 
cause for the subject's performance would probably include something besides 
the form of the schema, such as knowledge of the word 'C' in her language, 
willingness to play the game and assent or dissent, and so forth. But, again, 
Peacocke doesn't
 need to define causal sufficiency for the subject's performance; he only wants 
to show that this performance, if it can be demonstrated, whatever its cause, 
is sufficient to indicate possession of the concept CONJUNCTION.
So, this little paragraph does not, by my lights, do much damage to Peacocke's 
Thanks for your patience!

--- On Thu, 9/30/10, Ron Allen <wavelets@xxxxxxxxxxx> wrote:

From: Ron Allen <wavelets@xxxxxxxxxxx>
Subject: Re: [quickphilosophy] Fodor on Concepts IV: Circularity + Peacocke
To: quickphilosophy@xxxxxxxxxxxxxxx
Date: Thursday, September 30, 2010, 10:36 AM


Hi qp group:
Ah, it seems that my first post on this topic did make it to the list, but my 
spam filter directed my email copy to the Spam folder. It protected me from 
spamming myself. I feel so much safer now.
Anyway, Fodor's target now is two camps that are fellow-travelers of the BCP 
crowd: the conceptual role group (Boghossian) and the implicit definition group 
(Peacocke). It seems to me that BCP is more limited in its claims that these 
other thinkers. In other words, you can sort and make inferences without 
necessarily having a set of rules for conceptual roles or an implicit 
definition. So Fodor seems to be attacking those that go a step beyond plain 
vanilla BCP.
Also, it would seem to me that Boghossian would have to support BCP's 
conditions for concept possession. If you were sophisticated enough to follow a 
set of rules for expressions, then you'd be able to distinguish items in the 
extension (sorting), and you'd be able to make inferences (which is just to 
follow the rules in various situations). Similarly, holding an implicit 
definition would imply being able to sort and being able to infer.
The upshot is that Fodor could damage conceptual role semantics and could 
flummox implicit definiton and yet leave vanilla BCP unscathed.
Where does Fodor's attack begin? First he observes that conceptual role 
semantics cannot be simply asserting that it's possible to provide rules. 
That's easy enough to do in the case of referring expressions, like 'bachelor'. 
What could be dicey for conceptual role advocates to claim is that grasping a 
concept is to be identified with grasping the rules for the introduction and 
elimination of the concept, as in the case of CONJUNCTION given in the last 
(The introduction of CONJUNCTION just means that you are given p and q and you 
deduce pCq, introducing the C-symbol. Elimination means that you are given pCq, 
and from that you infer p (eliminating the C-symbol) or q (no C again).)
Fodor states that in the case of the logical constants, that's what conceptual 
role semantics has to do. In other words, Fodor is saying that in the case of 
their treatment of the logical constants, that conceptual role semantics merges 
with implicit definition concept possession. I think that's what he's saying 
in the 3rd paragraph beginning on p. 41 of his article.
Fodor is going to argue that the stronger claim to which the conceptual role 
and implicit definition camps must commit--namely that introduction and 
elimination rules must be identified with concept possession, not just possible 
from concept possession--cannot be sustained.
I hope this isn't too plodding. I find this part of Fodor's article difficult.

--- On Wed, 9/29/10, Ron Allen <wavelets@xxxxxxxxxxx> wrote:

From: Ron Allen <wavelets@xxxxxxxxxxx>
Subject: Re: [quickphilosophy] Fodor on Concepts IV: Circularity + Peacocke
To: quickphilosophy@xxxxxxxxxxxxxxx
Date: Wednesday, September 29, 2010, 6:45 AM


Hi quickphilosophy group: 

I tried to post this last night, but it didn't stick for some reason. Probably 
my ISP at fault. 

Fodor has another argument in Section 3.3 of his paper 'Having Concepts' that 
argues against 'conceptual role' semantics and the idea of an implicit 
definition as the basis for concept possession. 

1. Conceptual role semantics assigns rules to expressions that govern their 
use. There might be rules for referring expressions, such as 'flywheel', but 
the best examples are of logical constants, such as AND and OR, because these 
don't seem to have referents, and the only way to explain how they are used is 
to specify use rules. An example of a conceptual role semanticist is Paul 
Boghossian, whose forthcoming paper Fodor cites. (Fodor writes in 2004, and 
Prof. Boghossian's paper appeared in Nous in Aug 1996; here's a link to a draft 
on his NYU 
He says that "Implicit definition: It is by arbitrarily stipulating that 
certain sentences of logic are to be true, or that certain inferences are to be 
valid, that we attach a meaning to the logical constants. More specifically, a 
particular constant means that logical object, if any, which makes valid a 
specified set of sentences and/or inferences involving it."

2. The implicit definition camp is also represented by Christopher Peacocke, (A 
Study of Concepts, Cambridge, MA: MIT Press, 1996). Peacocke says that 
"Conjunction is that concept C to posess which a thinker must find transitions 
that are instances of the following forms primitively compelling, and must do 
so because they are of these forms:

q      pCq      pCq
---     -----      ------
pCq  p          q      "

(I hope that format aligns well in the group's page.) This is on p. 6 of 
Peacocke's book. Fodor inserts "[inferential]" before "transitions" in the 
quote above. So, basically, you possess the concept of CONJUNCTION if and only 
if you're compelled by these transitions, which define the notion of logical 

Let me see if this reply sticks.


--- On Mon, 9/13/10, Ron Allen <wavelets@xxxxxxxxxxx> wrote:

From: Ron Allen <wavelets@xxxxxxxxxxx>
Subject: Re: [quickphilosophy] Fodor on Concepts IV: Circularity
To: quickphilosophy@xxxxxxxxxxxxxxx
Date: Monday, September 13, 2010, 10:47 PM


Hi kwikphil group: 

Just to recapitulate: Fodor finds in section 3.3 of his paper 'Having 
Concepts,' that there is a vicious circle in the BCP analysis of concept 
possession. The sorting criterion (one of two, the other being inferential 
skill) that is partly sufficient for concept possession depends on the notion 
of the concept possessor possessing the same or an equivalent concept in order 
to perform the sorting. So, attempting to explain concept possession in terms 
of sorting and inference relies on (by way of the first condition, sorting) a 
notion of concept possession already, and the concept in question is the same 
or conceptually the same ans the one that BCP hopes to analyze. Circular, and 
viciously so at that.

But Fodor goes on to remark (p. 40) that this has irked many philosophers, and 
Wittgenstein is one of them. Here is how Fodor reviews W's difficulties in 
being a sophisticated BCP advocate, beset by the burden of the circularity that 
Fodor is gesturing at:

1. For W, having a concept involves knowing a criterion for applying a concept.

[rla, insert] Note that a "criterion" for Wittgenstein is a problematic and 
plastic term. I think that Baker and Hacker comment upon this, and that along 
with Fodor, we need to be cautious in throwing the concept around.

2. Thus, minimally, one can apply the concept in good instances in favorable 

3. Second for W, the application criteria for a concept are constitutive of its 

[rla] PI.208 seems to say this.

4. Fodor picks an example again: PAIN. The criteria for its 2nd person 
application are behavioral.

5. To have the PAIN concept is to sort pains from non-pains [and (rla) to make 
inferences; BCP isn't asserting that sorting alone suffices] by behavioral 

6. But, the PAIN concept isn't the same as the BEHAVIOR-MANIFESTATION-OF-PAIN 

7. A good instance of the behavioral manifestation an instance of pain, or not?

8. Fodor does not think so, and he thinks that W should be confounded by this 
as well: No wonder that W said things like "pain isn't a something" but "it 
isn't a nothing either".

[rla: this is Philosophical Investigations 304].

9. In summary, Fodor claims that "...either there is no sorting condition on 
concept possession or, if you prefer, the sorting condition that there is is 
circular." He has also convicted Wittgenstein, a bona fide sophisticated BCPer, 
along with the hapless and crude BCP advocates (Skinner and Quine), of the 
charge of circular reasoning.

Next on the menu, Fodor is going to find vicious circularity within the 
inference condition that BCP asserts for concept possession. I'll cover this in 
the next few posts. Mostly, it's an attack on the theory of concepts that 
Peacocke advances in A Study of Concepts, MIT Press, 1992.


--- On Mon, 9/13/10, wavelets@xxxxxxxxxxx <wavelets@xxxxxxxxxxx> wrote:

From: wavelets@xxxxxxxxxxx <wavelets@xxxxxxxxxxx>
Subject: [quickphilosophy] Fodor on Concepts IV: Circularity
To: quickphilosophy@xxxxxxxxxxxxxxx
Date: Monday, September 13, 2010, 11:46 AM


Hi quickphilosophers:

Fodor claims that neither of the conditions that Bare-Bones Concept Pragmatism 
(BCP) finds to be jointly sufficient for concept possession, sorting and 
inference, can be formulated without a vicious circularity.

Beginning with the sorting criterion in Section 3.3 of his article 'Having 
Concepts,' Fodor worries about the extension of a sort. He proceeds by example. 

1. Consider someone that can sort DOGs according to the concept of a BARKER. 
This sort works as long as the only things that bark are DOGs and all DOGs 

2. So, for the sort to be a criterion for concept possession, it seems that 
sort must be successful in counterfactual conditions: the sort must work for 
cases of barking non-dogs, canine non-barkers, and so on.

3. Yet, it seems that even under counterfactual conditions, the extension that 
results from a sort does not determine the concept that motivates the sort.

4. To
 show this, Fodor offers more examples: (a) TRIANGLE and CLOSED-TRILATERAL, (b) 
TWO and EVEN-PRIME, and (c) WATER and H2O. 

5. You can have the first of these concepts, without having the second.

6. Thus, the necessary coextension of two sorts does not imply that the sorts 
reveal the same concept.

7. One must sort according to some concept that one has.

8. BCP must therefore claim that only sorts according to a conceptually 
equivalent concept can produce the same sort.

9. So BCP must claim that the sorting condition for a concept possession C 
presupposed some possessed concept C' that is either identical to C or 
conceptually equivalent to C.

10. This is a vicious circle. 

Fodor goes on to comment on how Wittgenstein attempted to deal with this 
circularity. I'll cover that in my next post.






Other related posts: