Are Dispositions to Believe Constitutive for Understanding?

ARE DISPOSITIONS TO BELIEVE CONSTITUTIVE FOR UNDERSTANDING?

Keywords:

Inferentialism, dispositionalism, massive modularity

Abstract

T. Williamson argues against the thesis he recognizes as one of the inferentialist basic idea that he formulates as understanding/assent link, the claim that the assent to a sentence (believing a thought, at conceptual level) is constitutive for understanding it. This paper aims to show that appropriately articulated dispositional theory, could plausibly account for a weak version of inferentialism.

Introduction

Among many topics Timothy Williamson challenges in his book “Philosophy of Philosophy”[1], the issue of the conditions needed for understanding words or grasping concepts has one of the central places. The question concerning this issue is: Are there constraints that as necessary (or even as necessary and sufficient) conditions determine the understanding of words, as constitutive parts of sentences, and of concepts, as parts of thoughts? The prevailing and rather commonsensical view is that there are such conditions[2]. The clearest cases in which understanding and its (putative) conditions are tightly connected are those concerning analytic sentences. Let us take simple examples:

Can one understand the word (or grasp the concept) bachelor, or the word (or the concept) vixen without assenting to the sentence

“Every bachelor is an unmarried adult male”, or

“Every vixen is a female fox”?

Or taking an inferential form, could we say for someone that she has or grasps the concept “and” without her being disposed to infer according to the simple inference patterns:

“a and b”, therefore “a” (“b”); or:

“a” and “b”, therefore “a and b”?

Tim Williamson, attacking the prevailing view, vigorously answers, YES[3]. Inferentialists, in contrast, take a kind of link expressed in the above examples to be a necessary condition for understanding such a concept as ‘bachelor’ or ‘vixen’ or a logical constant, such as “and”. I agree with the general spirit of inferentialism (claiming that there are some conditions), trying to develop a naturalistic variant of inferentialism, and disagree with Williamson with regard to this variant.

Inferentialism criticised

In his (2007) book but also in other papers (“Blind Reasoning”[4] (BR) and “Conceptual Truth”[5] (CT)) T. Williamson argues against the thesis he recognizes as one of the inferentialist[6]basic idea that he formulates as understanding/assent link, the claim thatthe assent to a sentence (believing a thought, at conceptual level) is constitutive for understanding it (prominent inferentialist’s names are,among others, Boghossian, Peacocke, Jackson and Brandom). Understanding/assent link is the claim that the assent to a sentence (believing a thought, at conceptual level) is constitutive for understanding it.

Williamson critically considers two variants of understanding/assent link. The stronger one, I dub it canonical, is put forward as:

(C) Necessarily, whoever understands the sentence S (or grasp the thought t), assents to it.

Regarding its canonical formulation, overall but accurately represented structure of this formulation might be depicted in this way:

Understanding S Assenting to S (disposition to assent)

The weaker one, for Williamson “watered down” formulation, I shall call dispositionalist account, is set up in this form:

(D) Necessarily, whoever understands the sentence “Every vixen is a vixen” (whoever grasps the thought everyvixen is a vixen), has a disposition to assent to it.

The formulation I shall defend in the paper is the weaker dispositional account that is nicely indicated byBoghossian, who, interestingly enough, only formulates, but not really endorses, let aside developsit, in this way:

“…the idea that,… , we come to grasp the logical constant by being disposed to engage in some inferences involving them and not in others, is an independent compelling idea. (Blind reasoning, 240).

Williamson forcefully argues against (C) offering a number of meticulous and compelling arguments but the central evidence against it is provided by counterexamples to (C) that give justice to a situation where one understands a sentence and yet dissents from it, rejecting in this way the understanding/assent link.

I completely agree with Williamson’s criticism of canonical (C) formulation, but not with his denial of dispositional (D formulation) account.

Concerning (D) formulation, the crucial claim is that the fact that one has a disposition to assent to the sentence does not imply that he assents to the sentence. The role of dispositions, being understood as sub-personal (as Dennett had introduced the term) in the sense of implicit, tacit, possibly unconscious processes constituting deductive competence, is actually dismissed as having any serious explanatory role.

This paper aims to show that appropriately articulated dispositional theory, opposing psychological and cognitive theories Williamson relies on, could plausibly account for a weak version of inferentialism according to which “…we come to grasp the logical constant by being disposed to engage in some inferences involving them and not in others”.

I want to develop the dispositionalist suggestion and defend it against Williamson's criticism.Concerning the question in the title of this paper, my answer is positive. I claim that normal reasoners have spontaneous, possiblyunconscious default dispositions to accept the simplest logically correct inferential moves (and, in the case of an analytic sentence, to assent to it). Having important explanatory role, these dispositions, I amgoing to claim, are constitutive for understanding a sentence or grasping a thought.

Let me present Williamson’s examples that show that inferentialist’s basic and clearest canonical statement:

Necessarily, whoever understands the sentence “Every vixen is a female fox”, assents to it (or has a disposition to assent to it),

does not hold. Williamson offers a tremendous amount of critical material. The central cases in denials of inferentialism that counterexamplifies understanding - assent linkare depicted as imaginary cases of Peter and Stephan. The third one has to do with logical principles (connectives). It mentions the logician Van McGee who,finding some instances of modus ponens invalid, dissents from the modus ponens rule.

Let me start with the case of Peter, the vixen sceptic. Peter has a weird belief that vixens do not exist, due to his enchantment with conspiracy theories. Peter also has a meta-linguistic theory that “every vixen is a vixen” logically entails “there is at least one vixen”, i.e. that existential quantifier contains ontological commitment (i.e. “Every F is a G” is true if (i) there is a value of the variable “x” for which “x is an F” is true and (ii) there is no value of the variable “x” for which “x is an F” is true while “x is G” is not (2006, p. 86). Due to these theories, Peter dissents from “every vixen is a female fox”, even from “every vixen is a vixen”.

The second example depicts the case of Stephan, the queer-vixen theorist. He believes that some clearly female evolutionary ancestors of foxes are borderline cases of the “fox” and, therefore,ofthe “vixen”. Concerning his meta-theoretical belief, “what worries him is vagueness. He believes that borderline cases for vague terms constitute the truth-value gap” He generalizes “two-valued semantics by treating the gap as a third value”. In the vixen-example, for the borderline case, “x is a vixen”, as a value of x, is neither true nor false. Accordingly, for Stephan, “every vixen is a vixen” is not true, although, it is not false.

Finally, Van McGee, the famous logician, states this examplein his article:

According to the opinion polls just before 1980 elections, Republican R. Reagan was ahead of the Democrat J. Carter, with the other Republican, J. Anderson as a distant third.

Those apprised of the poll results had good reasons to believe the nested conditional:

-If a Republican wins, then if it’s not Reagan who wins it will be Anderson.

-A Republican will win the race.

Yet they did not have reason to believe:

-If it is not Reagan who wins, it will be Anderson.

The issue is not here whether some instances of MP are invalid or not, and, accordingly, whether McGee is right in dissenting from those instances of MP. The point is that he finds some instances unacceptable and regards MP invalid. In this case he, on reflection dissents from MP and certainly understands it.

Let me summarize all three examples. The condition common to all of them is that:

(1) Subjects do not assent to the sentence (inference) in which the ingredient term occurs and yet understand the sentence.

Additional conditions are:

(1a) Considering 1) as true, it is necessary to suppose that “deviant” speakers, understanding the word (or grasping the thought) have the same concept of it as the ordinary speaker (thinker) does.

(1b) The mental attitude of assenting implies first order consideration, not the other way around (“Assent is no metalinguistic or metaconceptual attitude: normally, in actively assenting to “Grass is green” one is saying or thinking that grass is green, not that the sentence “Grass is green” or thought grass is green is true” p. 75).

I am in complete agreement with Williamsonregarding general condition 1, where he compellingly confirms the claim according to which subjects can understand the sentence and yet do not assent to it. I am ready to accept additional condition 1a as well. Regarding the additional condition 1b, it is part of my argument that it does not hold. I claim that default dispositions to believe a proposition (or assent to the sentence) are constitutive for understanding it up to the point wherea strong defeater might override the original disposition. The fact that something can override a disposition does not mean that the constitutive role of the original disposition is annihilated. Using Williamson’s counterexamples to (C), I am claiming that defeatersin the form of the conscious, reflective decision to dissent from the sentence, previously accepted as a default disposition, usually stem from meta-semantic or meta-linguistic level.

Optimistic view

The defender of the inferentialism, who, like me endorses dispositional account, insisting on the claim that understanding-assent link is set up on sub-personal dispositional level, can offer a more optimistic reading of the dispositional account.

Let me first introduce a few general notes on the notion of dispositions. Very widely discussed in different topics and understood in many different ways, a disposition as a psychological concept, as a reasoning mechanism, could be articulated in two ways, as taking place either on the sub-personal or on the personal level. At the sub-personal level, the postulated dispositional reasoning mechanisms are processes “in the brain”, possibly, as Williamson takes it to be, “grounded in something like an unconscious reasoning module” PoP.100. They are unconscious in the way that the reasoner is not aware of the processes itself, and they are implicit, tacit and intuitive in the sense that the reasoner holds its outcome, a belief, without awareness of reasons to have it. At the personal level they could be understood as reflective-based, reasons-tracking inclination or incentive to behave in a certain way (attraction to believe).

Williamson’s strategy against (D) formulation

In (D), disposition to assent to a sentence S (disposition A to S) is a necessary condition for understanding S. Dispositions are understood as sub-personal (as Dennett had introduced the term) in the sense of implicit, tacit, possibly unconscious processes constituting deductive competence.

Williamson’s denying strategy is here as follows:

1) If one accepts the (putative) fact that one has a disposition A to S, the counterexamples for (C) provide strong evidence that they can be overridden by a reflective and conscious dissent from S.

2) Dispositions, Williamson argues, are not only overridden, they are also annihilated or erased. “It is certainly unclear how having a concept could consist in having a disposition that is overridden by one’s conscious reflection?” BR, p. 254

It can be deduced from (1) and (2) that dispositions do not play any important explanatory role, they cannot explain anything, particularly, sub-personal dispositions cannot explain reflective considerations at a personal level.

The claims 1) and 2) are enriched by the third claim. It states that:

3) Even if sub-persona dispositions have some role in human cognitive economy, the gap between sub-personal processing and personal beliefs is too wide to be bypassed.

Relying on recent psychological literature, Williamson decisively claims:

“thus the evidence suggests that the unconscious logic in question is not at the service of the cognitive processes that normally produce conscious assent to sentences like “Every vixen is a vixen”.” PoP,109.

Hence, the two steps argument, claiming that, first, overridden dispositions are also erased and, second, even if they play some role in cognitive economy, they cannot influence reflective and conscious processes. Therefore,

DA A link

My objections to Williamson’s position will go against both steps of “anti-dispositionalist” account. Answering the claim 2) I shall discussWilliamson’s counterexamples trying to show that they are not as lethal as it might seem.

Addressing claim 3) I will make use of the psychological evidence supporting the idea that sub-personal dispositionalist account can explain and cover full continuum starting form simple perceptual beliefs and going up to conceptual inferences. Those beliefs formation processes and inferences are implicit, tacit and intuitive in the sense of being output of the relevant competence. The entire continuum, from simple beliefs to complex inferences, up to reflective inferences, is smooth enough to avoid dramatic explanatory gaps. According to this picture, the reflective acceptance of an inference at a personal level is a tip of an iceberg,the body of which consists of different intuitive, implicit, tacit beliefs and inferences.

Contrary to Williamson’s understanding of assent as reflective and conscious, I claim that normal reasoners have spontaneous, possible unconscious default dispositions to accept the simplest logically correct inferential moves (and, in the case of analytic sentence, to assent to it). I endorse a view, more usual in recent literature that beliefs, and its sentential equivalents, assents, are to be understood in a dispositional-functional way. This theoretical approach claims that a belief is identified functionally, by determining its functional role. The functional role of a belief is its contribution, in practical reasoning together with other mental states like desired, to some kind of behaviour (mental or physical). Therefore, a belief is a disposition to behave appropriately. The propositions that are the object of beliefs could be entertained consciously or not. The most usual way of entertaining them is tacit, unconscious, implicit and spontaneous. In this way a belief (and assent as a relation to the sentence) is a disposition to behave as if a content of a belief (or a sentence) were true.

Coming back to our topic, to understanding/assent link (or to grasping/inferring link in Boghossian’s formulation), this approach to the belief (assent) theory provides a good ground for claiming that, in the case of an analytic sentences (natural world terms), grasping a concept one is disposed to infer according to inference pattern in the way as above.

Do dispositions really matter?

Williamson admits that Peter and Stephen “At some deep level, … have a disposition to accept (1) as true”, but, he continues “That a disposition is prevented from manifesting itself by a conscious reflection at an overlying level of theory-construction, …”. CT, 17. It seems inevitable to conclude, Williamson suggests, that sub-personal dispositions do not matter.

Dispositions do matter

To present the opposite view, a possible analogy with a process of perception might be helpful. Here we normally consider perceptual beliefs as factual and justified (in this case it is a kind of prima facie knowledge). We are normally disposed to form a great number of perceptual beliefs without being particularly aware of the way they occur or without reflecting about the reasons we have for possessing them.

In the case of perceptual illusions, for instance, when an unusual situation occurs, a perceiver tends to employ higher, sceptical epistemic criteria. He is now ready to readjust his default beliefs to make a broader world picture coherent. A perceiver can begin to doubt in, or dissent from the ordinary perceptual output, from the meta-theoretical perspective. A reflection or a personal level consideration might enter the process and usually does itas a meta-level theorizing (or pre-theorizing) about the process and its output. The final result is dissenting from the initial belief.

Does this imply that default, sub-personal (intuitive, being an outcome from deductive competence) processes have no explanatory value? My answer is no. Such processes, and the dispositions involved, do make a valuable contribution to the explanation of perceptual beliefs. To generalize the claim, default dispositions, both illusionary as well as veridical ones, are of indispensable and unavoidable importance for any kind of cognitive investigations. Hence, dispositions survive being overridden. The analogy suggests that dispositions, overridden by interfering conscious reflections, are not erased. They are still there and they can accomplish a lot of explanatory work. Furthermore, if we can account for the defeater and its role independently, treating it as an interfering element, and may suppose that (and how) the defeater itself can be defeated, and iteratively so, it turns out that the role of the sub-personal disposition is more than sufficient to rely upon.

Explaining the dissenters

But, what about dissenters Williamson so convincingly depicted to us? Let us start with Peter, the vixen sceptic example. Peter, as we are told, dissents from the elementary logical truth such as “every vixen is a vixen” partly because he does not believe that vixens exist. Furthermore, he declines to accept any sentence containing empty quasi-singular term “vixen”. Namely, most of his other beliefs Peter forms and changes implicitly and under the influence of many, possibly queer theories of the world. Also, he implicitly (intuitively) checks the validity of his inferences. Now, he has a belief:

“There are no foxes (and no vixens)”. He is still far from having the belief:

“ ’Every vixen is a vixen’ is not true”.

To dissent from “every vixen is a vixen” he must have a particular meta-semantic theory. This meta-semantic view enables him to consistently deny the property of identity to the “vixen” and dissent form the sentence “every vixen is a vixen”. In this sense, his dissenting is dissenting from the semantic form of the sentence, and only secondarily about vixen’s properties.