Tisztelt kollegak, az ELTE BTK Logika tanszekenek szeminariuman, november 26-an, hetfon 1/4 7-kor a Bp. V. Pesti B. u. 1. felem. 23. szobaban megvitatjuk Frakas Katalin mellekelt cikket. Erdeklodoket szivesen latunk. mate andras ============================================================================== Andras Mate CSc, assoc. prof. Inst. Philosophy, Dept. Logic Lorand Eotvos University Budapest, Faculty of Arts and Humanities H-1364 Budapest, POB 107 Phone: (36 1) 266 9100/5328 -- Phone/Fax/Answ: (36 1) 266 4195 e-mail: andras.mate(a)elte.hu Home: H-1152 Budapest, Kinizsi u. 132. / Phone: (36 1) 306 0584 {\rtf1\ansi \deff4\deflang1033{\fonttbl{\f4\froman\fcharset238\fprq2 Times New Roman CE;}}{\colortbl;\red0\green0\blue0;\red0\green0\blue255;\red0\green255\blue255;\red0\green255\blue0;\red255\green0\blue255; \red255\green0\blue0;\red255\green255\blue0;\red255\green255\blue255;\red0\green0\blue128;\red0\green128\blue128;\red0\green128\blue0;\red128\green0\blue128;\red128\green0\blue0;\red128\green128\blue0;\red128\green128\blue128;\red192\green192\blue192;} {\stylesheet{\f4\lang1038 \snext0 Normal;}{\*\cs10 \additive Default Paragraph Font;}{\s15\tqc\tx4536\tqr\tx9072 \f4\lang1038 \sbasedon0\snext15 footer;}{\*\cs16 \additive\sbasedon10 page number;}}{\info{\author M\'e1t\'e9 Andr\'e1s} {\operator M\'e1t\'e9 Andr\'e1s}{\creatim\yr2001\mo10\dy10\hr14\min38}{\revtim\yr2001\mo10\dy10\hr14\min41}{\printim\yr2001\mo10\dy10\hr14\min41}{\version2}{\edmins2}{\nofpages0}{\nofwords0}{\nofchars0}{\vern49205}} \paperw11906\paperh16838\margl1417\margr1417\margt1417\margb1417 \deftab708\widowctrl\ftnbj\aenddoc\hyphhotz425\formshade \fet0\sectd \linex0\headery709\footery709\colsx709\endnhere {\footer \pard\plain \s15\tqc\tx4536\tqr\tx9072\pvpara\phmrg\posxc\posy0 \f4\lang1038 {\field{\*\fldinst {\cs16 PAGE }}{\fldrslt {\cs16 7}}}{\cs16 \par }\pard \s15\tqc\tx4536\tqr\tx9072 \par }{\*\pnseclvl1\pnucrm\pnstart1\pnindent720\pnhang{\pntxta .}}{\*\pnseclvl2\pnucltr\pnstart1\pnindent720\pnhang{\pntxta .}}{\*\pnseclvl3\pndec\pnstart1\pnindent720\pnhang{\pntxta .}}{\*\pnseclvl4\pnlcltr\pnstart1\pnindent720\pnhang{\pntxta )}}{\*\pnseclvl5 \pndec\pnstart1\pnindent720\pnhang{\pntxtb (}{\pntxta )}}{\*\pnseclvl6\pnlcltr\pnstart1\pnindent720\pnhang{\pntxtb (}{\pntxta )}}{\*\pnseclvl7\pnlcrm\pnstart1\pnindent720\pnhang{\pntxtb (}{\pntxta )}}{\*\pnseclvl8\pnlcltr\pnstart1\pnindent720\pnhang {\pntxtb (}{\pntxta )}}{\*\pnseclvl9\pnlcrm\pnstart1\pnindent720\pnhang{\pntxtb (}{\pntxta )}}\pard\plain \f4\lang1038 Sense determines reference \par \pard Katalin Farkas \par Central European University, Budapest \par \par 1. \par \pard In his famous paper, "On Sense and Reference" Frege introduces the distinction between the sense and the reference of a name. The reference of a name is a definite object, and the sense is "wherein the mode of presentation is contained". As to the relation between a sign, its sense and its reference, Frege says that "to the sign there corresponds a definite sense and to that in turn a definite reference". This idea became k nown in the tradition influenced by Frege as the doctrine that "sense determines reference". The doctrine that sense determines reference has been interpreted in various ways; I shall assume here a simple and specific understanding of determination. S det ermines R will mean simply that sameness of S implies sameness of R. \par \pard Frege extended the sense/reference distinction to sentences as well. In an oft-cited passage of "Sense and reference" Frege first states that a declarative sentence contains a thought, and then inquires whether the thought should be regarded as the sense or the reference of the sentence. The reference of the sentence should not change if we replace a name in the sentence by another having the same reference. Yet we find that the thought can change by such an operation: "The morning star is a body illuminated by the Sun" and "The evening star is a body illuminated by the Sun" express different thoughts. The evidence for this is that someone might hold the first thought to be true, the other f alse. So thoughts can be objects of cognitive attitudes, and one criterion for their difference is this: if T1 and T2 are thoughts, and it is possible for someone to hold conflicting cognitive attitudes towards them (e.g. believe one and not believe the o ther), then T1 and T2 cannot be the same thought.\'02 Further well known considerations lead Frege to the view that the sense of a declarative sentence is a thought, and the reference of the sentence is a truth-value. The doctrine mentioned earlier - that sens e determines reference - holds also in the case of thoughts. \par \pard Frege returns to the question of thoughts in a later paper, "The thought". A thought is regarded as something for which the question of truth can arise.\'02 Frege still holds the determination relat ion between sense and reference: it's not only that thoughts are true or false, they also determine their truth-value. The other characteristics of thoughts, that they can be objects of cognitive attitudes is also discussed. Towards the end of the paper, Frege explains how thoughts effect our life: someone grasps a thought, holds it to be true and the thought will influence her actions. \par \pard Thus there are two features Frege attributes to thoughts: that they should be truth-evaluable (truth-value determining) an d action-relevant Of course, this double role is not a mere coincidence. If the thought is relevant to our actions through its being grasped and held true, it would be really odd if the same thing would not also be something for which the question of trut h arises. \par \pard \par 2. \par \pard In contemporary philosophy of mind, mental contents (the successors of Fregean thoughts) are expected to perform the same tasks: determine a truth-value and be explanatory of our actions. \par \pard \tab There have been, however, a number of objections against the idea that a single notion of content could satisfy both requirements. One problem for any single-notion approach is supposed to be generated by indexical thoughts \endash thoughts whose truth-value depends on the context of the thinking of the thought. For example if Rudolph Lingens thinks it is raining here, this might cause him to open his umbrella. But it seems that the thought expressed merely as \'93 it is raining here\rdblquote is not sufficient to determine a truth-value: we also need the value of \'93here\rdblquote to b e filled in by the context. So supposing that Rudolph Lingens is standing on Trafalgar Square while thinking the thought, what is usually regarded as truth-evaluable can also be expressed by saying that it is raining on Trafalgar Square. But believing thi s does not have the same immediate link to action as believing that it is raining here does. Or put it in another way, it seems that these two thoughts do not satisfy the criterion of difference mentioned above: for it is possible for Rudolph Lingens to b elieve that it is raining here without believing that it is raining on Trafalgar Square.\'02 Hence truth-evaluable and action-relevant contents are not always the same. \par \pard \par 3. \par \pard Exploiting the phenomenon of indexicality appears in an interesting way in Putnam's argument for semantic externalism - the view that meanings are individuated by facts external to a subject. In "The Meaning of 'Meaning'"\'02 Putnam claims that the following two assumptions are incompatible: \par \pard (I) \'93that knowing the meaning of a term is just a matter of being in a certain psychological state\rdblquote (i.e. narrow psychological state) \par (II) \'93that the meaning of a term (in the sense of \rquote intension\rquote ) determines its extension (in the sense that sameness of intension entails sameness of extension)\rdblquote (219) \par \pard \tab The two assumptions being incompatible, one of them has to be given up (222). Which assumption is to be given up depends on the type of expression we are considering. For indexicals, Putnam claims, \'93 no one has ever suggested the traditional theory that \lquote intension determines extension\rquote \rdblquote (234)\'02. What about the term \'93water\rdblquote ? The basic intuition in the Twin Earth thought experiment is supposed to be that \'93water \rdblquote refers exclusively to H2O on Earth and exclusively to XYZ on Twin Earth. Water, on Putnam\rquote s analysis, is wh atever which bears the same-liquid relation to this stuff (pointing to an instance of water), and "this stuff" picks out H2O on Earth and XYZ on Twin Earth. On page 234. of "The Meaning of 'Meaning'" Putnam says: "Our theory can be summarised as saying th at words like 'water' have an unnoticed indexical component". There are two ways, Putnam goes on, to incorporate the insight that \'93water\rdblquote has an indexical component. One would be treating it on an exact analogy with other indexicals. On this view, \'93water\rdblquote w ould have the same meaning but different extension on Earth and Twin Earth; that is, we retain the first assumption and reject the second, and apparently no externalist conclusion follows. The other way to incorporate the indexical insight would be to sti ck to the doctrine of the second assumption, and say that "difference in extension is ipso facto a difference in intension" (ibid.). This option is chosen by Putnam and it leads straight to the semantic externalist conclusion. \par \pard Putnam's original argument was about meanings, but the thesis of externalism assumed its real significance when it was extended to mental contents. Externalism about mental content is the view that the content of (some) of our mental states is individuate d by facts external to us. Here is a possible argument from the indexicality of \'93water\rdblquote to content externalism. Suppose that Leo Peter believes that athletes drink four litres of water every day. As I said, content is expected to be truth-evaluable. In order to evaluate the truth of Leo Peter\rquote s belief, the relevant value provided by the context must be filled in: that is, the content must somehow include a specification of whether athletes drink H2O or XYZ, according to Leo Peter. Now it turns out that truth-evaluable content is different for Leo Peter and his Twin-Earth counterpart, which means that it cannot be internally individuated. \par \pard \par 4. \par \pard The idea that content should be truth-evaluable was exploited in a similar fashion in the Rudolph Lingens example. And that story is of course the same as Putnam's: internalism is incompatible with determining reference. As Paul Boghossian puts the moral of the Twin Earth thought experiment: "either those concepts don't determine what they refer to in some context-independent way (they are not individua ted in terms of their referents) or they do determine what they refer to so are not individuated individualistically" (my emphasis).\'02 \par \pard We should all agree that the identity of contents must somehow turn on their truth-conditions. Now consider an ordinary indexical like \'93here\rdblquote . We cannot say that \'93here\rdblquote in \'93it is raining here\rdblquote does not make any contribution to the truth-conditions of the sentence. Of course it does -- replace "here" with "there" and you get a sentence with different truth-conditions. The con tribution these words make is such that the truth-conditions of the sentence will be context-dependent. The assumption in the externalist argument is that no such thing is allowed for "water", and similarly, no such thing is allowed in the case of mental contents. This is a basic assumption in all arguments for content externalism: \par \pard \par (E) Mental contents have, or are individuated by, context-independent truth-conditions. \par \par \pard If we rejected this assumption, then it seems that we would not be forced to adopt exter nalism even if we accepted the basic intuition of the Twin Earth thought experiment. For then we could insist first, that if water has an indexical component, then it should be treated like an indexical, and hence as having the same meaning on Earth and T win Earth. No semantic externalist conclusion follows. Second, we could maintain that the content of an Earthling's water thoughts have context-dependent truth-conditions, the same as the context-dependent truth-conditions of the thoughts of her Twin Eart h counterpart. Since we rejected (E), we do not have supply context-independent truth-conditions. No externalism about mental content. \par \pard So what obstructs the development of this line of internalist response is the view that meaning determines reference (in the way understood by Putnam), or in other words, that contents have context-independent truth-conditions. In what follows, I would li ke to probe this assumption. \par \pard \par 5. \par \pard So what is the origin of the assumption (E)? And what is its justification? When Putnam introduces the idea that intension determines extension, he says: \'93 ... it was taken to be obvious that (...) two terms cannot differ in extension and have the same intension. Interestingly, no argument for this impossibility was ever offered.\rdblquote (op. cit. 218 ) And Putnam himself doesn't do anything to improve this state of affairs. He speculates that the assumption probably reflects the medieval tradition of holding that a concept was a conjunction of predicates which give necessary and sufficient conditions for something falling into the term\rquote s extension. Bur surely, this cannot be Putnam\rquote s reason for accepting the doctrine. \par \pard It will be natural to look for the origin of the doctrine in Frege's writings. Frege is also well known for the view that an indexical sentence does not, in itself, express a complete thought. Let's then consider one of the passages where he says this. \par \pard \'85 is the thought changeable or is it timeless? The thought we express by the Pythagorean theorem is surely timeless, eternal, unchangeable. But are there not thoughts which are true today but false in six months time? The thought, for example, that the tree is covered with green leaves, will surely be false in six months time. No, for it is not the same thought at all. The words 'this tree is covered with green leaves' are not sufficient by themselves for the utterance, the time of utterance is involved as well. Without the time-indication this gives we have no complete thought, i.e. no thought at all. But this thought, if it is true, is true not only today or tomorrow but timelessly. ("The Thought" p. 103) \par \pard \par \pard Notice that the putative objection against the unchangeability of a thought was that it may change its truth-value. The only reason to think that the thought expressed by "The tree is covered with green leaves" has changed between now and six months time is that it has changed its truth-value. Once this objection is cleared away - by claiming that it is not the same thought which has different truth-value - Frege draws the conclusion that tho ughts are indeed unchangeable. Only the inessential properties of thoughts may change - for example the property of being grasped by me or by someone else. \par \pard This shows that according to Frege, thoughts have their truth-value essentially. If two sentences differ in truth-value, they cannot express the same thought. I said above that the claim that thoughts determine their truth-value is an instance of the doct rine that sense determines reference. This, we can see now, is quite literally true: sense alone determines reference. \par \pard \par 6. \par \pard The doctrine seems plausible in the case of mathematics and logic. Given that the Pythagorean theorem is true, no false sentence could express the same thought as the Pythagorean theorem does. If Frege had these kinds of examples in mind in the first plac e, it is easy to understand why he adopted the doctrine. But consider a contingent sentence like \'93The inventor of bifocals was a man\rdblquote . As it happens, the description picks out Benjamin Franklin, who was indeed a man. So the sentence is true. Now as far as I know, Frege doesn\rquote t discuss questions arising in connection with alternative possibilities; but such questions are often raised in contemporary philosophy of language or mind. If thoughts have their truth-value essentially, then it is impossible to express the same thought by a false sentence. Hence in a world where Mrs Franklin and not Mr Franklin invented bifocals, the sentence could not possibly express the same thought. \par \pard But, contrary to this, it is standard to assume that in a wo rld where this sentence is false, its meaning would nonetheless be the same. Same holds for the corresponding belief and its content: the belief could be false and have the same content. If this is right, then we cannot say that meaning (or content) alone determines a truth-value; we also need the state of the world \endash what we might call the circumstances of evaluation \endash that is, in this case, the fact that the inventor bifocals was a man. (An analogous reasoning holds for the description "the inventor of bi focals": its meaning alone is not sufficient to determine its extension. We also need the world to make its contribution.) So when we say that sense (or meaning) determines reference, we understand this as relative to some circumstances of evaluations. \par \pard It seems then that considerations about other possible worlds suggest a departure from the conception that sense alone determines reference.\'02 Alternatively, someone could propose the following reasoning. We do want sense to determine reference (content to det ermine truth-value), but we see that sense understood as above is not sufficient in itself to determine reference. So we need to introduce another notion of sense or content, one that \endash for this sentence - determines reference properly. Now compare Herber t Garner in this world and his counterfactual counterpart who lives in a world where Mrs Franklin invented bifocals. Despite the internal sameness of Herbert Garner and his counterpart, this newly introduced content of their belief that the inventor of bi focals was a man will be different. Consequently, content is externally individuated. \par \pard In the case of true sentences, this newly introduced content is going to be something like a fact (a true sentence). If the content of Herbert Garner's belief is the fact that the inventor of bifocals was a man, then clearly the corresponding belief of He rbert Garner's counterpart cannot have the same content: for in that world, there is no such fact that the inventor of bifocals was a man. It is somewhat more difficult to s ay what kind of thing the content is in this case, but this is not surprising: externalism is always biased towards semantic success. Compare a case with indexicals. In the original example about Leo Peter, it was easy to provide the truth-evaluable conte nt of Leo Peter's belief that it is raining here, for "here" referred to Trafalgar Square in the context. But consider another example. Leo Peter is having a conversation with Gustav Lauben. Leo Peter says: "You are wounded" According to the now familiar li ne of reasoning, the truth-evaluable content of this can be expressed as "Gustav Lauben is wounded". But suppose further that in another situation, Leo Peter merely hallucinates that he is talking to Gustav Lauben, and in fact no-one is there. In this cas e, it will be difficult to state what the truth-evaluable content of "You are wounded" would be, given that "you" refers to nothing. Some externalists may suggest that in such a situation, Leo Peter is in fact not thinking a thought. These externalists pu nish failure of semantic success with refusing to grant thoughts. But it would be extremely implausible to extend this theory to our newly introduced contents: no-one would want to say that only true thoughts can be thought.\'02 \par \pard \par 7. \par \pard I don\rquote t see much sense in introducing these new kinds of contents, and I don\rquote t think anyone would want to argue for externalism in this way. The fact that purely descriptive thoughts\rquote truth or falsity is partly externally determined (by the way the world is) has never been regard ed as an argument for externalism proper. So it seems that we can compromise on the idea that reference is determined entirely by something internal without thereby committing ourselves to externalism in any interesting sense. \par My question is then this: why not pursue the same strategy with indexicals?\'02 The characteristic feature of indexical thoughts is that, besides the circumstances of evaluation, we need further factors to determine their truth-value. Now I don't want to suggest that the distinction betw een circumstances of evaluation and contextual features should be obliterated; we might need the distinction for various semantic or other purposes\'02. But suppose that someone is going through the following step-by-step reasoning. \par \pard We start with mathematical or logical statements. It is quite plausible that the thought, whatever we think, whatever we have in mind when grasping a mathematical statement is sufficient in itself to determine a truth-value.\'02 \par \pard Then we consider contingent statements. We recognise th at the situation here is somewhat different: there is a factor relevant to the determination of truth-value which was not present in the previous case. To put the matter in a much simplified way, we have to make a choice whether this further factor should belong to the world or to the individual mind. The choice is between these two options: shall we give up the idea that thought alone determines reference, keep thought the same as it was, and trust the external world to supply the missing determination, OR, shall we pack the further determining factor into the thought. The decision here is almost incontrovertible: circumstances of evaluation belong to the world and not to the mental realm. \par \pard Next we consider indexical statements. Here we encounter a further type of truth-value determining factors. I claim that it is up to us to make a choice again: do these factors belong to the individual mind, to what can be characterised as the realm of th inking - or are they external to it? Suppose that someone had independent reasons to opt for the second alternative: for example that these factors appear to be further detached from what is relevant to our actions.\'02 Then we make the same decision as before. We acknowledge the existence of these newly discovered truth-value determining factors, but we place them in the world rather than in the individual mind. What is wrong with this idea as such? Nothing. \par \pard Someone could try to justify opting for the other option by citing the doctrine that sense determines reference. To th is my reply is that we have already given up the idea that sense alone determines reference, without, however, committing ourselves to the view that sense has nothing to do with reference. We just do a bit more of the same. \par \pard We started with the idea that thoughts have their truth-value essentially, and we had to abandon this idea. Instead, what is customary to say these days is that thoughts or contents have their truth-conditions essentially. Sometimes people speak as if ins isting on internalism even in the light of Twin Earth thought experiments would mean giving up the very idea of the truth-conditionality of mental content.\'02 This is surely unfair. As I said above, indexicals do make a contribution to truth-conditions: the sentences "I am wounded" and " you are wounded" have different truth-conditions. These truth-conditions are context-dependent. But there doesn't seem to be anything inherent in the notion of truth-conditions which would forbid this. Given the central role of the assumption that content h as context-independent truth-conditions, one should like to see some weighty arguments supporting it. These arguments would have to uncover an essential difference between circumstances of evaluation and contextual features - the difference which explains why should we treat the latter, as opposed to the former, as entering to the individuation of mental content. In the absence of such argument we are not forced to accept the assumption. \par \pard \par 8. \par \pard Summing up. It is customary to hold that the moral of the Twin Ear th thought experiment is that either we accept content externalism, or we give up the very idea of the truth-conditionality of content. I argued that this choice is not forced upon us. We can accept the basic intuition of the Twin Earth thought experiment , we can hold on to the idea that content is individuated in terms of truth-condition, and it is still open to us to resist the externalist conclusion. I would not want to say that I presented an argument against externalism, in the sense that I have show n that anything is wrong with externalism. My argument was perhaps not even a positive argument for internalism. All it was meant to do is to remove an impediment from the development of an internalist theory of content. \par \pard \par \'02 Evans' Intuitive Criterion of Difference, Varieties of Reference 18ff \par \pard \'02 \'93Without wishing to give a definition, I call a thought something for which the question of truth arises\rdblquote p. 87 of Gottlob Frege 1918: \'93The Thought\rdblquote in Simon Blackburn & Keith Simmons 1999 (eds.): Truth Oxford Univ ersity Press \par \pard \'02 Or so it seems. Some philosophers hold that these beliefs are the same. One thing seems to me clear: whether this is right or wrong, it is a departure from Frege's conception. \par \'02 Mind, Language and Reality Cambridge University Press 1975. All references to Putnam are to this paper. \par \pard \'02 This is probably incorrect. Evans' theory of the sense of demonstratives seems to be a case in point. Of course, this simply means that Evans was an externalist not only about natural kind terms, but also about demonstratives. \par \pard \'02 P. 164 of "What the Externalist Can Know A Priori" Proceedings of the Aristotelian Society XCVII/2 (1997): 161-75 \par \pard \'02 How far this is a departure from Frege, I cannot judge. Since Frege is silent on the question of alternative possibilities, it is difficult to say what would be the most natural extension of his theory to these questions. \par \pard \'02 The only case where this might be plausible is the case of God, supposing that God would not even entertain false thoughts. In fact, the present suggestio n is not unlike Leibniz's idea about the complete concept of an individual, which, however, can be thought only by God. Needless to say, my interest here lies in human thinking. \par \'02 This move would probably go against the Fregean conception. I am not sure how to situate these question with respect to Frege's ideas, since I am interested in questions arising about private psychological matters, which is manifestly different from Fre ge's acknowledged concerns. \par \pard \'02 David Kaplan, for example, has argued that to do so would be surely mistaken. See his "Demonstratives" section VII. pp. 507-10 in: J. Almog & J. Perry & H. Wettstein (eds.): Themes from Kaplan. Oxford University Press 1989: \par \pard \'02 I would like to remind the reader of what I said above in footnote 3 about the "unloaded" nature of the determination relation. "Determined by" should certainly not be read as anything like "depends on". Then it seems that what I say here does not imply any non-factualism or conventionalism or mind-dependence about mathematics. \par \pard \'02 This is, of course, a very heavily debated issue, and I cannot enter into the question here. \par \pard \'02 A sligthly different formulation from the one quoted above in another paper by Boghossian: "The minimal intended significance of the Twin Earth thought experiments may be put like this: Either mental content are not individuated individualistically, or they are not individuated in terms of their truth-conditions." "The transparency of mental content" Philosophical Perspectives 8 (1994): p.34 \par \pard \par \par \par \par \par }