Random thoughts on random topics

Traboule, Lyon

Certainties, truth and relativism

If we compare the picture that we can extract from On Certainty with Davidson’s view (as expounded in, e.g., ‘A Coherence Theory of Truth and Knowledge’), the important difference seems to be this, that Wittgenstein introduces the layer of certainties in between our epistemological practices and external reality, whereas Davidson construes the relation between belief and reality much more directly. The fact that certainties are categorically different from beliefs and other epistemological entities (despite the fact that over time, and between communities and/or individuals,  what counts as what may change) in combination with the plurality of systems of certainties, makes room for a measure of (conceptual) relativism that Davidson seeks to avoid. His way of doing so is to take the core of our belief system to be as stable (over time, over communities and/or individuals) as is the causal influence of external reality on humans. (There is more room for differences in the ‘superstructure’ of complex beliefs that are not directly caused by our interactions with reality, but that is something that Davidson does not pay that much attention to). 

This has also consequences for how truth works in both perspectives. In On Certainty truth is first and foremost a concept that operates within a particular epistemological practice, that itself is made possible by a particular system of reference consisting of certainties. (That Wittgenstein construes it in more or less verificationistic terms is an additional,  independently motivated feature.) The relation between external reality and certainties is not one of determination, but of constraint. This is the source of plurality, and it also implies that certainties are not upheld because they are true. The fact that different certainties  can be upheld at the same time also testifies to that, of course. Nevertheless, certainties differ in terms of their entrenchment and some are so basic to our form of life that it does not seem that much of a stretch to call them ‘true’, admittedly in quite a different sense. In Davidson’s perspective we also have two distinct properties. ‘Mild correspondence’ is the notion of truth that links beliefs (and hence meaning) to the world. It is what  the causal influence of the world on us results in. Internally, i.e., within our actual epistemological practice, truth then takes on a different form, that of coherence. 

Martin Stokhof
from: Aantekeningen/Notes
date: 22/03/2012

Random thoughts on random topics

Phoenix Mountain, Beijing

On language as the medium of hermeneutic experience

Gadamer, in Truth and Method: “Interpretation […] is the act of understanding itself, which is realized—not just for the one for whom one is interpreting but also for the interpreter himself—in the explicitness of verbal interpretation.” This is the claim that language is the medium of hermeneutic experience. (And all experience, of whatever kind, is also hermeneutic.) Gadamer insists that the absence of explicit linguistic formulations does not constitute a counterexample: all non-linguistic demonstrations in fact rely on language. 

The general claim, that all understanding is linguistic, i.e., has language as its medium, definitely sounds counterintuitive, since we do intuitively feel that there are things that we understand but that we can not ‘put into words’. 

One thing to bear in mind, tough, is the intimate relation between understanding and interpretation: all understanding is interpretation, and all interpretation results in understanding. To the extent that this sound wrong, it might reflect a hidden assumption about the existence of something like ‘ultimate’, ‘final’, ‘true and complete’ understanding, a kind of understanding that transcends the kind of understanding that interpretation results in. That, Gadamer claims, is an illusion. 

That means that from a Gadamerian point of view our problem in fact reduces to the question: ‘Is all interpretation verbal (linguistic)?’ Again, we need to take that in a broad sense, i.e., without any presupposition of actual verbalisations. (In an analogous fashion, Gadamer argues that we need not worry about linguistic diversity.)

The claim then seems to come down to this: anything that is proposed as an interpretation is in principle subject to questioning, argumentation, justification. This must be so, for interpretation itself is the result of such questioning, etcetera. Asking questions, providing answers, disputing and justifying them, all this is done in language (or in a medium that presupposes language), and inasmuch as there is something that fails to be subject to these language-based procedures it can not part of the interpretation itself (and hence can not constitute (part of) understanding). 

It’s not that Gadamer would deny the existence of non-verbalisable phenomena, the claim is that as such they are not part of interpretation and hence not part of our understanding of something.  (Note that this comes remarkably close to Wittgenstein’s analysis of the ‘tip-of-the-tongue’ phenomenon in Philosophical Investigations, II.xi.)

To what is extent is this a feasible position? It seems it all centers around the question whether we can actually point to the existence of a kind of understanding that meets two requirements: it is essentially non-linguistic; it is somehow connected to the kind of understanding that is linguistic.

The second requirement is the really problematic one, I think. But it also seems justified. For without it, the dispute would in fact be merely verbal: Is there something that is not linguistic understanding? Of course there is. Can we call it ‘understanding’? Well, yes, but what would be gained by that? It is only when we can point to relationships between the two phenomena, that we really confront Gadamer’s position.

So the question is: Can we do that? What would be good examples?

Martin Stokhof
from: Radical Discussion Board
date: fall 2004

Random thoughts on random topics

Zollverein, Essen

Metaphor

The following is a curious catch in Davidson’s account of metaphor. If a metaphor has only literal meaning, then what exactly does it mean to say that a speaker uses a metaphor? What a speaker uses, in any case, is a sentence, and, according to Davidson, any sentence only has literal meaning. So, ‘metaphor’ as it is used here cannot be a predicate of sentences, but at best expresses a property of utterances of sentences, i.e., of use. ‘To use a metaphor’ then must mean: ‘to use a sentence metaphorically’. But is that not at odds with Davidson’s insistence that metaphor is not a matter of speaker’s intention?

Martin Stokhof
from: Interpretation
date: fall 1998

Random thoughts on random topics

Quiberville

Charity and false beliefs

What Charity does is to assume agreement (in beliefs). That by itself does not automatically assume truth (of those beliefs), it does so only in conjunction with the assumption that we ourselves have the right beliefs. And there is the conundrum: we know that some of our beliefs are false, but we do no know which these are (otherwise we would not hold them). So indirectly we will assume that others, even if we assume they agree with us, will not be right all the time, but for the same reason that we can not attribute false beliefs to ourselves without giving them up, we can not attribute false beliefs to other without giving up the assumption of agreement. 

Martin Stokhof
from: Interpretation
date: fall 2008

Ramblings on random topics

Sotteville-sur-mer

On interpreting rooks

‘Rooks console each other.’ (NRC Handelsblad, 23/01/2007) Question: can a rook console me? Can I console a rook? If in both cases the answer is ‘No’, then we have two isolated subframes, each characterised by a ‘console’-relation that has certain (formal and material) properties within that subframe, but which is limited to that subframe. The analogy with (in)translatability is obvious. The question that arises then is this: What reasons do we have (could we have) to call two such completely isolated relations both an ‘x-relation’? In what sense are these two the same relation?

Martin Stokhof
from: Aantekeningen/Notes
date: 23/01/2007

Random thoughts on random topics

Saint Petersburg

On interpretation

What is the mystery of interpretation is not that it would be impossible: we are not able to fly, nor do we have the ability of bi-location, and no-one wonders why we don’t. Nor that it would be: we breath, walk, and again we take this for granted (which does not mean that we can not investigate the actual physiology of these abilities). Regarding interpretation the real problem seems to be that it is both possible and impossible. That interpretation comes natural to us — understanding somehow happens to us without us knowing how —, and that it is an impossible task — no matter how much conscious effort we put in to it we are in the end always defeated. The otherness of the other makes us strangers to ourselves, yet our familiarity with ourselves makes us understand them. Conscious and mechanical, possible and impossible: interpretation is always all of that, but none of it completely.

Martin Stokhof
from: Interpretation
date: 05-1997

Random thoughts on random topics

Berlin

On transitivity of translation

Part of the argumentation in Davidson’s ‘On the Very Idea of a Conceptual Scheme’ centers around transitivity of the translation relation. The point Davidson is trying to make here, I think, is that if there exists a translation from L1 into L2, that translation itself can be stated in either L1 or L2, and consists in an accurate mapping of sentences from L1 onto sentences of L2. If the translation is stated in L1 then, by assumption, it can also be stated in L2 (after all, there is a mapping from all of L1 into L2, i.e., also of the L1-sentences describing the mapping). So we can safely assume that if there is a translation from L1 into L2, it can be formulated in L2. Now suppose there is a translation from L2 into L3: that, again is a mapping, in this case of L2-sentences onto L3-sentences. That mapping includes those L2-sentences that describe how to translate L1-sentences into L2-sentences. So what we have then is (among other things) a set of L3-sentences that tell us which L2-sentences are translations of which L1-sentences, and L3-translations of the L2-sentences. But that means we have a translation of L1 into L3. (More neatly: if a translation is a homomorphism from L to L’, then we know that if there is a homomorphism from L1 to L2 and a homomorphism from L2 to L3, then there is a homomorphism from L1 to L3, viz.: the composition of the two.)

Of course the above argument only works if we assume that translations are total, i.e., that they map all sentences of L onto sentences of L’ and vice versa. What if we drop that assumption? First of all we have to ask ourselves whether we are still dealing with translation in such a case. But let that pass, and suppose we have a mapping of all the sentences of L1 onto a proper subset of sentences of L2, i.e., there are parts of L2 that have no counterpart in L1. Notice that the formulation of the translation can not be in the latter set (for that would mean that the translation would be statable in L2, but not in L1, which is absurd.) Now assume we have a translation from L2 into L3: the only way in which that would not give us a translation of L1 into L3 (by the reasoning above), is by being restricted to exactly that part of L2 that does not contain the L1-to-L2 translation. But that would mean that it really isn’t a translation of L2 as such: it must leave out a proper part of L2. Of course, such a mapping could exists, but we would lack any justification for calling it a translation. (And if we would insist that we could call it that, then we actually are assuming what we set out to prove, viz., the failure of transitivity of translation.)

Note that the fact that in actual cases there are bound to be discrepancies between languages, i.e., things in L that don’t have an exact counterpart in L’, or things in L’ that lack a counterpart in L, does not really affect this line of argumentation. The central point is that the translation itself concerns such a substantial part of both languages, that that by itself guarantees transitivity to a sufficient degree.

Martin Stokhof
from: Radical Interpretation Discussion Board
date: 11-2004

Random thoughts on random topics

Lisbon

Uninterpreted content

Theodore Schatzki’s analysis of dispersed and integrative practises implies that normativity arises, not at the higher order level of rules, or practises, or institutions (at least not exclusively), but at the very basic level of individual interaction with the environment. This view is reinforced by various analyses of know-how and expertise.

What is interesting to note, especially with regard to Davidson and Gadamer, who actually are in the same boat here, is that uninterpreted content then plays a key role. We need such content in order for normativity to have a basis on which the language-based practises may build.

For Gadamer this is anathema: all experience is linguistic and exists only in and through language. For Davidson it presents a problem too, although it may not be immediately obvious that it does. For doesn’t Davidson avail himself of a primitive causal relationship between the world and us? And doesn’t he reject any form of mediation (linguistic or otherwise) between ourselves and the world?

But unmediated content is not the same as uninterpreted content as we use the phrase here. For in Davidson’s view our causal interaction with the world results in beliefs, and beliefs and (sentence) meanings are indistinguishable in terms of structure and content. So Davidson’s unmediated content is highly structured, in such a way that it is immediately expressible (given a suitably expressive language, of course): definitely not the uninterpreted content of everyday expertise. In fact Davidson seems committed to the same kind of linguistic view on experience that Gadamer embraces explicitly.

Martin Stokhof
from: Aantekeningen/Notes
date: 21/11/2006

Random thoughts on random topics

Amsterdam

Davidson, indeterminacy and charity

The point seems to be this: given that we have a shared ontology due to the application of the Tarskian framework (as the theoretical framework in which we formulate concrete theories of meanings for concrete languages) and charity, which implies shared beliefs, why don’t we have a shared vocabulary and a shared theory of reference concerning this vocabulary? 

The point seems to be this: given that we have a shared ontology due to the application of the Tarskian framework (as the theoretical framework in which we formulate concrete theories of meanings for concrete languages) and charity, which implies shared beliefs, why don’t we have a shared vocabulary and a shared theory of reference concerning this vocabulary? 

The question boils down to what follows from the assumptions mentioned about reference. Let’s start with the use of the Tarskian framework. For Davidson this follows from the assumptions he makes concerning the nature of meaning (extensionalism) and the function of a semantic theory (explanation of competence); cf., ‘Truth and Meaning’ for the details. So his view is that if we want a theory of meaning for a language, it has to have that particular form. And that in its turn presupposes that in any language we find a shared logical machinery, consisting of propositional connectives, quantificational apparatus (and the basic distinctions brought along by that) and the logical rules governing their behaviour. But that is all, in particular it does not involve any assumptions concerning the reference of the non-logical vocabulary. Of course, reference is used in stating the Tarskian truth theory, but there it is used only as an auxiliary notion. What the theory defines, or accounts for, is (our knowledge of) truth conditions of sentences. It does so using the auxiliary notion of reference of sub-sentential expressions, but, and this is the important point: it does not define truth on the basis of an independent account of reference. And as Davidson points out, any account of truth along these lines leaves reference essentially underdetermined (a point also made by Putnam and several others): two sentences may have the same truth conditions under different assignments of references to their sub-sentential expressions. (If the sentences are from different languages, we have two translatable sentences which do not allow us to infer any shared reference; if the sentences are from the same language, this means that synonymy does not guarantee unique reference; and as a special (but most important) case we have that it is possible to assign to one and the same sentence the same truth conditions based on attributions of different references to one or more of its component expressions.)

Then charity. Notice that charity, too, concerns sentences, not words. In interpretation, i.e., in actually trying to construct a Tarskian theory of truth for a given language, the empirical data we start from are utterances made in a context (situation). The assumption that truth plays the same role for the speakers of the language we are interpreting as it does for us, these data can be viewed as utterances of sentences held true in that situation. A specific (to be determined!) subset of those will be utterances of sentences held true about that situation, i.e., utterances of sentences that are held true on the basis of certain aspects of the situation in which they are uttered. Each and every sentence uttered is supposed to express a belief. So certain sentences uttered in a situation express beliefs of the speaker about that situation. This is where charity comes in: it allows us to proceed on the assumption that the belief that the speaker expresses in these sentences (but remember, we do not know of every sentence in advance whether it belongs to this set) is a belief that we hold about the situation as well. This is supposed to give us enough common ground to work our way into the language. But, and this is the important point, just as two sentences can have the same truth conditions, yet not share reference of sub-sentential expressions, beliefs too can be shared without a shared set of objects, properties and relations that can be attached in a unique way as references to the expressions that occur in the sentences that are used to express these beliefs.

So neither the Tarskian framework nor charity allows us to venture beyond the level of sentences/beliefs and be confident that we will return with a unique ontology in this particular sense. But for Davidson the conclusion is not that hence there is a relativity of ontological schemes, but rather that the idea behind it, viz., that the beliefs we hold and the meanings of the sentences we use to express those beliefs, are built up from referents in this particular way, is misguided in the first place. In combination with the holistic nature of language and belief that Davidson clearly endorses, this assumption would lead to relativism. But there is, according to Davidson, no reason to make it in the first place.

Martin Stokhof
[from: Radical Interpretation Discussion Board
date: 10-2003]

Random thoughts on random topics

Tang pottery, Tsinghua University museum

Concerning Gadamer and normativity

The following seems a very plausible conjecture: it is the meaning of the text itself that provides the necessary normative constraints on its interpretation. But there are a few problems with that. 

First of all, it makes interpretation very much a factual, ‘realistic’ concern: independent from interpretations and interpreters, there is such a thing as ‘the meaning’, and the task of interpretation is to discover that. Once we’ve done that, the task is fulfilled and there is no more need for interpretation. But that doesn’t sit very well with Gadamer’s insistence that interpretation is an on-going affair, and moreover, one that not only constantly changes the views of the interpreter, but also the meaning(s) of what is interpreted: the ‘fusion of horizons’ is a temporary equilibrium, brought about by adjusting both the perspective of the interpreter and that of the text.

Secondly, if the objective meaning of the text itself were to play this role, this wouldn’t fit into an interpretational scheme that follows the hermeneutic circle. Recall that if we follow the structure of the hermeneutic circle we need to compare two things that both are different from this postulated objective meaning of the text itself, viz., the fore-projection, i.e., our ‘initial hypothesis’, and the result of our (first) reading. The problem was that we can compare these two without any problem, but that in order to evaluate the outcome of that comparison, we need a standard, something normative. Now suppose the objective meaning were to play that role? How would that help? If we know that this is the objective meaning of the text, we wouldn’t need any interpretation to begin with. And if we do not, it will fail to hold any normative authority.

The essence of the problem is that the hermeneutic circle, precisely because it is a circle, involves only entities of the same kind (meanings). And without reference to any external source of normativity, none of these can play the required normative role, on pain of the entire circular structure collapsing into what is basically a realistically understood concept of objectivity.

Martin Stokhof
[from: Radical Interpretation Discussion Board
date: 11-2006]