brandom/phil-2420/downloads/Week 6... · Web viewfollowing argument in mind: we start with practical deontic attitudes (normative attitude) which are manifest in our practices (nondiscursive

October 2, 2012

Week Six Questions

In chapter 4 of Making it Explicit, Brandom introduces perception as a“language-entry move” in the game of giving and asking for reasons. Empiricalcontent, although itself standing in inferential relations, is the result ofnoninferential elicitation. Observation, Brandom argues serves to function as aregress-stopper or is foundational since it is given default entitlement status (invirtue of the way in which the report tokening is elicited through the exercise of areliable differential response disposition). Although it seems as though the notionof observation as justified without justification stops the regress of entitlementto premises (since one’s entitlement to an empirical claim can be taken forgranted) it does not appear to stop the regress of entitlement to inference.

Brandom argues that we can end the regress of inference by “taking or treatinginferences as correct in practice” (205). He seems to have something like thefollowing argument in mind: we start with practical deontic attitudes (normativeattitude) which are manifest in our practices (nondiscursive norms implicit inpractice). In other words, we take something to be correct according to a practice.It is these pragmatic deontic attitudes which institute deontic statuses (the statusof being correct according to a practice). However, there seems to be a problem withthis supposed regress-stopper. Our practical deontic attitudes are themselvesessentially normative. Insofar as these attitudes function as an assessment ofcorrectness, they must already have a normative status. Since normative statuses areinstituted by normative attitudes, it seems that we require a further normativeattitude to explain the initial normative attitude, presumably this would continueon infinitely. This regress could possibly be circumvented by accepting primitiverules of inference (as Brandom later argues we should). But these primitive rulescannot, for the reasons just recounted, be derived from norms implicit in practice.It remains to be seen where these primitive rules have their origin.

Laura Davis

Here is a question about the optionality of acknowledging material inferences in theorder of being versus in the order of explanation:

In a characteristically permissive tone, Brandom says (Articulating Reasons, p. 86)that it is optional whether or not one accepts material inferences as goodinferences independently of the formal criterion of validity. It is clear that hischoice to accord explanatory primacy to material inferences over formal ones isfundamental to the explanatory strategy running through his philosophy. Yet I wonderwhether there is an analogous but mandatory ontological (metaphysical) primacyindependent of this optional explanatory primacy between the two types of transitionone can make between claims (or claims and actions).

For the way in which the regress of rules argument motivates the project of MakingIt Explicit essentially involves constraints not only on what counts as a feasibleapproach to the topic of conceptual contentfulness, but also regarding theconditions for there being a normative space of reasons at all. It specifically rules out the option of understanding concepts as rules of reasoning

that are explicitly formulated as statements of rules, because the normative forceof concepts could not exist under such conditions alone.

Now, in holding that material inferences are but enthymemes, is not a formalistrelying precisely on this ruled-out conception of reasoning? The "missing" premisesin material inferences, the formalist maintains, must be made explicit at the outsetor else one drawing the inference is not following rules. On the general conceptionof reasoning as acting on the basis of conception of rules, then, such moves betweenclaims (or claims and actions) are not rational. Against this train of thought,however, it seems that one, including Brandom, ought to say that it is not optionalwhether or not one allows for material inferences in reasoning, because the regressof rules argument shows that without such implicit transitions between claims (orclaims and actions) there cannot be reasoning at all in the more demanding sense ofrule-following.

In other words, does not the regress argument work as a transcendental argument foraccepting material inferences? Seen from this perspective, the status of materialinferences emerges as an ontological (or metaphysical) question regarding adependence between two kinds of transition one can make in reasoning. Unlike in thecase of variously expedient strategies in the order of explanation, here, addressingthe order of being, is not acknowledging material inferences not optional butindispensable?

P.S. To add some historical depth to this question, one could raise the same issueby asking how Sellars understood the connection between the position he defended in"Inference and Meaning" and the regress argument he expounded at the outset of "SomeReflections on Language-Games". How, after all, does he *defend and motivate* thenotion of material inference in the paper? As you undoubtedly can see, there aremany threads to elaborate here, but I suppose I should leave them for anotheroccasion.Tuomo Tiisala

MIE, chap. 4:

In the former part of the chapter 4 of MIE, Brandom points out an interesting parallel structure between two important paradoxes in semantics and epistemology, that is, Kripkenstein’s rule-following consideration (discussed in the chapter 1) and the traditional problem concerning the foundations of empirical knowledge. In both cases, there are two familiar types of approach: (1) Regulism in semantics and justificatory internalism (of non-Sellarsian kind) in epistemology; (2) Regularism in semantics and reliabilism in epistemology. According to Brandom, each approach in (1) suffers a sort of regress problem, while each in (2) suffers a sort of “gerrymandering” problem. This structural parallelism between the semantic and epistemic paradoxes is not so surprising, once we recall that in each of them, an analysis of a normative status, though of different kind, is at stake --- that is, the correctness of concept-application (associated with the concept of truth) in the former, and the entitlement (the “J” condition of the JTB model of knowledge) in the latter.

Impressed with this parallelism, I would like to pursue some further consequences of it. Let us focus on the gerrymandering problem for the second type of approach (i.e., regularism and reliabilism). According to Brandom’s diagnosis (pp. 211-2), the culprit of this problem is that the second type approaches are naturalistic and thereby too conceptually austere to make the required distinctions: Regularists/reliabilists cannot distinguish the privileged regularity/reference class from an infinite number of irrelevant others, and thus gerrymandering follows. Contrastively, Brandom claims that his normative pragmatism does not suffer the same problem, since he helps himself to the normative concepts that enable him to make the distinctions required. Having said this, I would like to raise two interrelated questions, aiming to compare naturalism with normative pragmatism similarly in semantics and epistemology.

(a) Is it really the case that naturalists cannot avoid the gerrymandering problem? Isn’t there any naturalist substitute that can simulate the behavior of the normative distinction at issue?

In the semantic side, one of the relevant resources at hand for naturalists is consideration of evolutional history. At least as far as a primitive level of “concept”-application concerns, the concept of proper function developed in teleosemantics, the function in virtue of which its bearer has been selected through the process of evolution, seems to do a good job in determining the boundary of such a primitive kind of “concept.” Although it is not so clear how such an account can be extended to the more sophisticated kind of concept-applications as we do, such a line of explanation still seems to be a possibility.

In the epistemic side, it seems that naturalists can pursue a similar line. For example, they might be able to specify the relevant reference class based on which the reliability of an organism’s behaviors should be assessed, in terms of something like a living sphere that the organism usually resides in and moves around. And this concept of living sphere might also be applicable to us, human being, though such an application would require a further sophistication of that concept. (Here, I temporarily set aside another more serious flow of reliabilism, which is discussed in pp. 214-5, that reliabilism hardly respects the propositional contentfulness of knowledge (i.e., the “B” condition of the JTB model of knowledge).)

(b) Has Brandom’s normative pragmatism really offered a fully satisfying answer to the gerrymandering problem by itself?

In the semantic side, I am worrying about how at most a finite number of normative attitudes concerning the past applications of a concept can finally institute the correct application condition of that concept that demarcates the infinitely many individual things to which that concept is potentially applicable. In other words, I wonder how normative pragmatists can stride over the gap between the finiteness of our practice and the infiniteness of the meaning it institutes, which is one of the essential sources of the gerrymandering problem. At this stage, I have a suspicion that normative pragmatists might need a help of such a modal concept as

disposition, which the teleosemanticists in (a) prefer to use. (I think this whole question is very relevant to the one Billy raised in the third lecture.)

Similarly in the epistemic side, normative pragmatists seem to leave it unanswered how the privileged sort of reference class can be chosen in each individual situation. The only thing that Brandom says about this issue is that it is a “pragmatic” matter, and therefore the answer can vary depending on “the interests and goals of those performing speech acts” (p. 213). Certainly, this view seems to be plausible. However, it does not seem to be sufficient, if necessary, to substantially answer the original question of how to pick out a particular reference class from those irrelevant.

Shuhei Shimamura

*My question is essentially a follow-up to questions I've asked before,which I hope can be made a bit more precise now that we've moved furtheralong. You've mentioned cases where two interpreter of a community mightdisagree about attributions of practices, such that they could even, inprinciple, disagree about whether or not a particular supposed instance ofthe practice was successful or not. I take it that there will even beinstances in which the interpreters disagree about whether or notthe community has properly assessed some instance of a practice asa successful or unsuccessful instance of that practice. I'm still unsurewhether or not you think these disagreements will be faultless or not, and,if they are, if there will always be exactly one correct interpretation. ****This line of thought leads me to wonder if two interpreters could(faultlessly) disagree about whether or not a given reporter in a communitywas reliable. I assume the use of scare quotes and would allow trivialcases to drop out (I have in mind cases where the two interpreters disagreeabout the reporter because one or the other is unwilling to undertakecertain substitutional commitments attributed to the community orreporter.) Is it a consequence of refusing to naturalize reliability thatthere may be cases where there is no fact of the matter about whether ornot a given reporter is reliable, at least to outsiders of the culture? Ifso -- is this a good thing?*

-Billy Eck

Question:Suppose my whole life I have been working the railroad tracks, trapped in awindowless perch just above them. When my boss tells me to move the trackto the left I simply push a button. I am told that if I don’t, then thetrain cannot run on the correct track; and if the train cannot run on thecorrect track, food cannot get to the town on time etc. So, when I am toldto move the track to the left, I understand the command in virtue ofunderstanding the inferential lattice it belongs to. Suppose that I nowescape my perch armed with RDRDs. *What do I learn when I expose my RDRDs

to the outside environment?* I think I learn what it is in virtue of whichI use and understand language as I do, and so learn what it is in virtue ofwhich the inferential relations among propositional contents hold. Inexposing my RDRD to the contours of the bit of track I have thus far neverobserved, I gain observational knowledge of what it is in virtue of whichthe track must be moved to the left, and not to the right etc. *What isthe status of this observational knowledge? *This cannot be a gain insemantic knowledge since, by hypothesis, I understood the languageperfectly well when I was cooped up in the windowless perch. It seems tome that what I now have is *meta-semantic* knowledge: I know the reasonthat the inferential lattice takes on the form it does. The lattice I hadonce understood *sola fide* is now, in some sense, grounded by the proximalupshots of RDRDs exposed to my outside environment. Brandom may want totake a different route: he may want to claim that when I was trapped in theperch the contents I understood were propositional, and only when I escapedthe perch did they slowly transform into something empirical (221). Onthat view, I undergo a shift in semantic knowledge: I understand thecontents differently. But the identity conditions of an empirical way ofunderstanding propositions are vague, e.g., what exactly controls the shiftfrom propositional to empirical contents?; how many observationally boundcontents must be connected to a content in order to make it empirical?;does this number depend on the contingencies of speakers? If, as I amrecommending, the knowledge I gain is meta-semantic, this would suggestthat in addition to logical vocabulary, the *RDRD* organ is one aspect ofthe “organon of semantic self-consciousness”. I am probably confused aboutsome part of your view here; if so, please disabuse.

Shivam Patel

*MIE, Ch. 4:*

In 4:V:3, you claim that normative vocabulary (“broadly evaluative words”),much like logical vocabulary, play in an expressive role by “makingexplicit material properties of…reasoning”, where the former makes explicitthose properties in *practical* reasoning and the latter reasoning in *general* (245ff). You exploit an analogy between the cases to justify yourclaim. First, you observe that:

Just because the endorsement of *p* → ¬*q* would bar the inference from *p *to*q* does not show that *p* → *q* is a suppressed or implicit premise in thematerial inference from *p *to *q* (248).

To gloss: We need not countenance that *p* → *q* plays a role in the abovereasoning, though we can appeal to it *to make explicit* the practices ofmaterial inference endorsed by a given player in a GOGAR who adopts*q*from her commitment that*p*.

You then claim that pro-attitude statements, which on some accounts areused “to bridge the gap between what one believes and what one decides todo” (247), can be treated much like *p* → *q* in the previous example. Tospell it out:

Just because the claim (*a*’) “I don’t want to keep dry” would bar theinference in chain of practical reasoning (a) “Only opening my umbrellawill keep me dry, so I shall open my umbrella” doesn’t mean that the claim (*a*) “I want to stay dry” is a suppressed or implicit premise in the chainof practical reasoning (a).

The idea here is that (a) is a completely valid material inference byitself. (*a*) can be appealed to *to make explicit* the practices ofmaterial inference endorsed by a given player in a GOGAR. Attributingclaim (*a*) to an agent “is just to take inferences” like the one in (a)“to be entitlement preserving” (249). In short, normative vocabulary, likelogical vocabulary, plays an expressive role in making explicit thematerial properties of reasoning, where the former makes explicit thoseproperties in cases of *practical* reasoning.

I’m inclined to think that these vocabularies are not in fact analogous,and that we cannot treat (*a*) in the case of practical reasoning asanalogous to *p *→ *q* in the case of general reasoning. A way to get atwhy I think so is to say that the reasoning in (a) is *not* a materiallygood inference without bringing in (*a*). Here are some good materialinferences for contrast: from “Berkeley is west of Pittsburgh” to“Pittsburgh is east of Berkeley”; from “This is copper” to “This conductselectricity”. Roughly, in both cases, communally shared inferentialarticulations of the concepts warrant the move. But nothing about thethose articulations in “Only my umbrella will keep me dry” warrants themove to “I shall open my umbrella.” Why should it follow from any beliefthat I shall (or should) do something?

Well, there are *some* beliefs from which it does follow that I shall (orshould) do something. These are evaluative beliefs, like “I want” or“ought to keep dry,” that express in their content why they are to *motivate* my action. If we accept this, then it seems to me that normative orbroadly evaluative notions must play more than an expressive role in thecase of practical reasoning. In fact, they play *the* central role. As Isee it, no reasoning can be rightfully called practical unless it purportsto preserves an evaluative element through material inferences. It’s truethat no suppressed conditionals need to be countenanced in cases ofpractical reasoning. But such reasoning always starts and ends with anevaluative judgment.

A caveat: Someone could perhaps object by saying that practical reasoningalways starts and ends with judgments containing “I shall…” rather than “Ishould…” This would be to deny that normative/evaluative language evenplays the expressive role of making explicit the material properties ofpractical reasoning. I think there are reasons to reject such a view infavor of a view that takes evaluation to be crucial to practical reasoning.This is a large discussion, but a brief response is that it is hard to seehow an action could belong to an agent without appealing to that agent’sevaluation of the action. Even some belief, intention or commitment tobring about a state of affairs that indeed causes that state of affairs tocome about seems insufficient to attribute and an *action* rather thanan *accidental

behavior* to an agent because of deviant causal chains considerations. Andif the conclusion of chain of practical reasoning contains an evaluativeelement, it’s hard to see how it could have gotten into the chain in thefirst place unless the chain began with an evaluative belief.

This discussion leads me to think that if the GOGAR is to have languageexits, then it must also contain normative vocabulary (or at leastconcepts). Normative vocabulary would then be more than merelyexplicative, but central to any autonomous discursive practice and alsothus to who we are.

Chuck Goldhaber