4. De re specifications of conceptual content The circumstances of production of a discursive text appealed to in justifying de dicto specifications of the contents of ascribed commitments provide only one important inferential context against the background of which to specify a claimable or believable (but not necessarily claimed or believed) conceptual content. The rules and elasticities I was worrying about above evidence the difficulty of codifying principles for specifying (from an at least somewhat different perspective) what someone thinks they are committing themselves to by what they say, what they in some sense intend to be committing themselves to, what they would take to be consequences of the claims they made. But besides the question of what one takes to follow from a claim one has made, there is the issue of what really follows from it. If I claim that this coin is made of copper, I am, whether I realize it or not, committing myself to its melting at 1083.4° C. Unless that claim is true, what I have said is not true either. To vary the example: if Henry Adams believed that the inventor of the lightning rod did not reside in Philadelphia, and if Ben Franklin in fact is the inventor of the lightning rod, then Henry Adams believed of Ben Franklin (as, we might want to say, the inventor of the lightning rod) that he did not reside in Philadelphia. This sort of characterization of the actual inferential content of the claim Henry Adams made is just what is wanted when one is assessing the truth of that claim. For if one has discovered not only that Ben Franklin did in fact invent the lightning rod, but also that he did reside in Philadelphia, then one has found out that what Henry Adams said is not true. One must specify the content of a claim correctly in order to assess its truth. If the right thing to say is that what Henry Adams said is not true, then we must be specifying its content correctly when we say that Henry Adams claimed of Benjamin Franklin that he did not reside in Philadelphia.
That ascription employs what we might call a “denotationally de re” specification of the content of the ascribed claim. The rules for such ascriptions are that
In this weak, merely denotational, sense, if Ortcutt believes that the shortest spy is a spy, and Rosa Kleb is the shortest spy, then although he may have no way of knowing it, Ortcutt believes of Rosa Kleb that she is a spy. For he believes something that is true if and only if Rosa Kleb is a spy. Denotational de re ascriptions specify conceptual content by saying what it is one is talking about, in the normative sense of which object one needs to investigate the properties and relations of in order to assess the truth of the claim in question. Thus I can cross the chasm created by the vast differences of belief separating me from the Zoroastrian priest and extract from his extravagant remarks information that I can use as premises for my own inferences if I can specify the content of his claim not only with the de dicto “He believes that Zoroaster is granting us his beatitude,” but also the de re “He believes of the sun and of shining that it is doing that.”
The belief specified by a denotationally de re ascription and that specified by a de dicto ascription are the same belief. It is just the belief that the inventor of the lightning rod did not reside in Philadelphia that is the belief of Benjamin Franklin that he did not reside in Philadelphia. The difference is in the context of collateral premises in which the claim is situated in order to assess its inferential significance. In the de dicto case, one draws the auxiliary hypotheses for the multipremise inferences involving the target claim from other commitments the one acknowledging the target commitment would acknowledge. This is the perspective from which one wants to specify the content of a commitment if one is interested in what other commitments the speaker/writer in question would acknowledge, or in what he would do to try to bring about various kinds of states of affairs. In the denotational de re case, one draws the auxiliary hypotheses for those multipremise inferences from the facts that determine what actually follows from what. That is to say that each ascriber draws those auxiliary hypotheses from the facts as she takes them to be; that is the best any of us can do. The ascriber’s commitments are the facts as she takes them to be. That is why when the ascriber is interested in truth, i.e. in what she herself should be committed to, what she should rely on as premises for further inferences of her own, she assesses the inferential significance of the ascribed claim from the inferential context provided by her own commitments regarding how things actually are with what the other one is (according to the ascriber) talking about.
Once again, the important thing to realize (a point that is explained and argued for in much greater detail in Chapter Eight of Making It Explicit) is that the de dicto ascription of a belief that (t) and the de re ascription of a belief of t’ that (it) are not ascriptions of different beliefs. They do not ascribe beliefs with different contents. Rather, they specify the single conceptual content of a single belief in two different ways, from two different perspectives, in two different contexts of auxiliary commitments. The significance of the presence of one sentence among the premises of a multipremise material inference—the difference its presence makes to what does and doesn’t follow from the rest—depends on what the rest of the premises are. So in this sense saying what does and doesn’t follow from a sentence must be at least implicitly relativized to a set of commitments that serve as the background against which one is going to assess the inferential significance of the claim in question. The choice of auxiliary hypotheses that distinguishes denotational de re specifications of the conceptual content of ascribed commitments has at least an equal claim to illuminate the commitment undertaken as does the choice of auxiliary hypotheses characteristic of de dicto specifications of conceptual content. If the Colonel orders his soldiers to cross the river within twenty four hours, he is, in effect, ordering them to do, within the general bounds of their authority, anything that is necessary, and something that is sufficient to bring about the ordered result. If achieving that result requires cutting down 60 trees (and doing that is within the bounds of their authority, or the Colonel’s) then in a real and practically important sense he has ordered them to cut down the trees, whether or not he has thought about the matter or even would accept that that is a consequence of his order.
It follows from this way of thinking about meaning that besides admitting de dicto intellectual historiography, we ought also to acknowledge the legitimacy of de re textual interpretations. These will be specifications of the very same conceptual contents that are specified by de dicto ascriptions. But in the de re case, those contents are specified from a different point of view: from the context provided by collateral premises that are, from the point of view of the ascriber, true. De re specifications of conceptual content attempt to say what really follows from the claims made, what is really evidence for or against them, and so what the author has really committed herself to, regardless of her opinion about the matter. The de re style of intellectual historiography requires laying facts alongside the claims of the text, in extracting consequences, assessing evidence, and so delineating their conceptual content. Responsibility for justifying these auxiliary hypotheses rests with the ascriber, rather than with one to whom the commitments whose contents are being specified are ascribed. So if Russell can establish that there are at least two things one can mean by “X is a part of Y”—one corresponding to set membership and the other to set inclusion—he is entirely justified in querying Plato to see what can be made of various of his claims when we distinguish the two senses. (The most devastating outcome would be to find that on occasion he uses the term “part” with the circumstances of application appropriate to one of the senses, but drawing consequences from that application that are appropriate only to the other.) If Sellars can establish that ‘experience’ can be used either to mean the act of experiencing something or the content that is experienced, then he is justified in interrogating Berkeley’s arguments to see which of them can be made out with one consistent interpretation of the term.
A generation ago the history of philosophy tended strongly toward de re readings. (One might think in this connection of the vastly influential works by which Strawson and Bennett for the first time made Kant’s theoretical philosophy into respectable topics for analytic philosophers.) If I read the sociology of the current situation correctly, there has been a substantial backlash to this practice, in favor of immensely patient and textually informed de dicto readings. I hope it is clear that I don’t think there is anything wrong with going about things in either of these ways. But it is a mistake to think that one or the other of these styles of content specification gets things right in a way the other doesn’t. Both are wholly legitimate ways of specifying the contents of the very same conceptual commitments expressed by the words on the page. It is only if one masquerades as the other, or is just unclear about the rules it acknowledges in selecting auxiliary hypotheses—that is, about the inferential context it is operating in—that error or confusion results. The response counseled by recognition of the essentially perspectival character of conceptual content construed as inferential role is irenic, tolerant, and pluralist: let a hundred flowers blossom.
And notice that in each case, once the context from which collateral premises are to be drawn has been specified, there can be an equally objective matter of fact concerning what the inferential significance of a textual claim is relative to that context. That is, de dicto and de re readings can each be rigorously assessed as to their correctness in specifying conceptual content relative to a context. We can disagree and make mistakes about, investigate, and resolve disputes concerning what actually follows from what is said, once a context is specified from which to draw our auxiliary hypotheses. And the same can be said for our inclusion of various claims in such a context, once the kind of context (de dicto or de re) has been settled. For one must justify the attribution of a given claim as one the author did or would acknowledge commitment to, that is, must justify taking it to be a licit collateral premise in the de dicto case. And the ascriber must justify each claim he takes to be true, that is, must justify taking it to be a licit collateral premise in the de re case. The fact that one can independently and individually assess the rational warrants for attributing (in the de dicto case) or endorsing (in the de re case) the claims that make up the inferential context with respect to which conceptual content is specified mean that these claims provide friction for and constraint on that process. In this respect they play a role in hermeneutic discourse analogous to that played by noninferential observation reports in empirical discourse. It should at any rate be clear that the relativity of specifications of conceptual content to inferential context as here construed in no way has as a consequence that “anything goes” or that the meaning of a particular text is wholly indeterminate or “up for grabs”.
We are now in a better position to understand why the distinction between extracting what is already a fully-formed inferential significance from a text, on the one hand, and foisting one on it from somewhere on the outside on the other, is unhelpful in thinking about the conceptual hermeneutic enterprise. Such applications of a making/finding distinction are inappropriate in light of the relativity of inferential significance to a context of collateral commitments. The conceptual content of a claim can in principle only be specified against the background of some such set of commitments. The interpreter has considerable choice in selecting such a context or inferential perspective. But once such a point of view has been selected—paradigmatically, once the choice has been made to offer a particular variety of de dicto or de re content specification, and so to privilege a particular inferential context—then it is not at all up to the ascriber what the significance of the claims in question is in the chosen context. The context is, if you like, made; but then the inferential significance of a text in that context is found. The perspectival character of conceptual content ensures that both moments, making and finding, will be in play in any ascription.
In this particular way, and for the special (but central) case of conceptual content, the inferentialist theory of meaning of Making It Explicit explicates and justifies the gadamerian denial that the making/finding distinction confronts the theorist of textual interpretation with a genuine dilemma. In this same sense (explication and justification for a central but special case), that theory can be seen to underwrite the other large scale hermeneutic claims I picked out earlier as “gadamerian platitudes.” Authorial intentions do play a privileged role in de dicto specifications of conceptual content. (Though of course, different sorts of de dicto ascription may assign this privilege differently: New Critical readings may restrict us to the words on the page, for some way of drawing the boundaries around the text considered. Others may allow into the context other texts or remarks of the same author, or even other things we can infer about her attitudes.) But this is because the claims the author acknowledges commitment to serve to define the contexts with respect to which a content specification counts as de dicto. It is not because those contexts are themselves privileged in that they provide specifications of conceptual content that are more correct, adequate, or true than de re specifications done from the point of view provided by other contexts. All contexts define in principle equally valid semantic perspectives from which to specify the conceptual content of a claim. Authorial intentions, whether stated by the author or inferred by an interpreter provide just one sort of context against the background of which inferential significance can be assessed. Any privileging of de dicto over de re ascriptions must be local and temporary, rooted in pragmatic, rather than semantic considerations—and here by ‘pragmatic’ I mean the vulgar sense of relativity to the purposes, interests, and plans of the interpreter.
This hermeneutic consequence of inferentialist semantic theory results from the perspectival character of contents, according to that semantic theory. And that is to say that the denial of certain kinds of authorial authority is a consequence of the analog within that theory of conceptual content of the gadamerian relativization of meaning to context, in a sense broad enough to include the commitments acknowledged by the interpreter, as well as those acknowledged by the producer of a text. What I’ve called the “perspectival character” of inferential roles (and hence of conceptual contents) is that the inferential significance of a claim—paradigmatically, what new consequential commitments result from undertaking such a commitment—is primarily a matter of its role in multipremise inferences. Since we have many choices concerning those collateral premises, each of which yields a genuine inferential significance of the claim, and so a genuine perspective on its inferential role, there are many contexts with respect to which its content can be specified in ascriptions. This conceptual perspectivism accordingly underwrites the interpretive pluralism that is another hallmark of gadamerian hermeneutics. For the same reason, at least in the specific case of conceptual content, the inferentialist approach to meaning offers a justification of the claim of the open-endedness of the sort of semantic interpretation one undertakes in specifying the content of a commitment one ascribes to another. Every new text provides a new context, against the background of which one might assess the inferential significance of any given claim.
5. Tradition and dialogue One context that is of particular significance is that of the tradition in which one situates a particular text. Establishing such a context—the sort of thing I try to do in Tales of the Mighty Dead—is itself no negligible accomplishment. As Eliot says in the essay quoted above, paradoxically: “Tradition…cannot be inherited. If you want it you must obtain it by great labour.” Here one supplements the words on the page by further claims made by others whom the interpreter, but not necessarily the authors involved, retrospectively sees as engaged in a common enterprise, as developing common thoughts or concepts. One might treat such ascriptions de traditione as another species, besides ascriptions de dicto and de re. I prefer to use “de re” generically, to refer to any ascription relative to a context (from a point of view) that is not restricted to commitments the interpreter takes it would be acknowledged by the author of the text—i.e. to use it as the complement to “de dicto”. The paradigmatic case, where the further commitments defining the inferential context are those acknowledged by the interpreter, can then be marked out as immediatede re ascriptions. If at least some of the collateral commitments appealed to in extracting inferential significances are ones the interpreter attributes, but does not acknowledge, then the de re specification of conceptual content can be said to be mediated by those attributions. In the important special case of ascriptions de traditione, the context is a mixed one. For delimiting a tradition involves both undertaking commitments concerning the relations of various texts one to another, and attributing commitments on the basis of what is said in those tradition-defining texts. The reason for adopting this generic use of “de re” is that in producing a specification of conceptual content from the point of view provided by any arbitrary context of collateral commitments, the interpreter must, among other things, do what he would do if those commitments were his own and he were making an immediate de re ascription. The interpreter must, in all but the de dicto cases, in this sense implicitly adopt the perspective from which the content-specification is being offered. As the mixed attitudes essential to ascriptions de traditione show, however, this is by no means all that can be going on in mediated de re ascriptions. Further structure of various sorts may also be significant. Because of their distinctive deontic structure, ascriptions de traditione are a particularly significant kind of mediated denotational de re ascription, and deserve their own designation.
One central and characteristic gadamerian trope presents interpreting a text as engaging in a kind of dialogue with it. This is a way of talking about a distinctive structure of reciprocal authority exercised by, and reciprocal responsibility incumbent upon, interpreter and interpreted. Once again, for the case of specifically conceptual interpretation, thinking about a context of collateral commitments as what relates conceptual content to inferential significance and thereby supplies the necessary background for specifications or characterizations of such contents in explicit ascriptions offers a dialogical model with a further articulated structure. It is useful here to take account of the difference between dialogical relations and dialogical processes, and think a bit about the relations between them. De re readings of any sort are inherently dialogical in a relational sense. First, they commingle premises from two different sources (voices, in an extended sense). This is one sense that could be given to Gadamer’s talk of a “fusion of horizons”. In this sense, each of them has its ‘say’. For the collaboration of the commitments of the two as it were interlocutors consists in their relation to their joint inferential consequences. The consequences they lead to are in general common in the sense that the support of each is required for the conclusion, rather than in the sense of being shared, i.e. already a consequence of what is drawn from each source. They are shared in the sense in which Fred and Ginger share a dance (something intelligible only in terms of what they are both doing), though they are moving differently, rather than in the sense in which soldiers marching in step share a gait. Something emerges inferentially from the collaboration of premises that was not contained in any of them apart from its fellows—though such consequences may be thought of as implicit already in the premise, in the perspectival sense that it would follow if the premise is set in the right context.
There are dialogical processes and practices in play, too. Interpretation in the sense of interpreting is a kind of doing. Even in the case of de dicto readings, the consequences of a set of premises must be extracted by the interpreter. In de re readings, by drawing conclusions from the text in the context, the interpreter is actively mediating between two sets of commitments. Text-and-context, on the one hand, and interpreter on the other, each have their distinctive roles. Still, the interpreter’s activity is responsible to the actual inferential relations. Except for the important case of immediate de re readings, then, the interpreter’s own commitments only make a difference to the outcome if she makes a mistake—if she doesn’t know or can’t figure out what really follows from what. (Of course we are often in that position. But that fact is not germane in the present context.) This is a consequence of the inferentialist semantic externalist claim that what really follows from what is not restricted to what is envisaged by the one having the belief or making the claim (or indeed, by anyone else).
The most important notion of hermeneutic dialogue underwritten by inferentialist semantics is a different one, however. For according to the development of that view in Making It Explicit, [is that] practical grasp or understanding of conceptual content is the ability to navigate and negotiate between the different perspectives from which such a content can be interpreted (implicitly) or specified (explicitly). This is the kind of know how that knowing, believing, or claiming that consists in. It is the capacity to move back and forth between the perspective-relative inferential significances made explicit in de dicto and de re specifications of one and the same conceptual content. When one can say both “S believes that a bunch of bloodthirsty fanatics occupied the village,” and “S believes of a bunch of gallant freedom fighters that they occupied the village,” one is calibrating claims (and concepts applied therein) according to the different doxastic perspectives of the author and the target of the ascriptions in a way that makes clear what inferential significance as premises they would have for each.13 Mapping different inferential significances, relative to distinct contexts, onto each other in this way is what taking them to be expressions of the same conceptual content consists in. For once again, it is the same conceptual content that is being attributed by the two ascriptions. Grasp of conceptual content in this sense is essentially dialogical, even in cases where one or more of the contexts in question is not associated with an interlocutor authorized to engage on its behalf in processes of expounding, expatiating, and answering for it.
This is the way mediated denotational de re ascriptions of the sort I have been calling de traditione are dialogical. The understanding they express involves “talking with a tradition” in a dual sense. One corresponds to an instrumental sense of ‘with’. An interpreter employs the tradition as a means of expression, a way of specifying the contents of, claims and texts it comprises. For one uses the commitments characteristic of the tradition as tools to extract from them an inferential significance—one perspective on a conceptual content. In the species of de traditione reading that are concerned with virtual semantic influence of the sort provided by any sort of context, rather than with actual causal influence, one may appeal to later developments in characterizing earlier ones. But there is also a conversational sense of ‘with’ in which one can talk with a tradition. The sort of understanding that is made explicit in immediate de re characterizations of the claims and texts a tradition comprises is a critical one. For it is manifested in the process of moving back and forth between the perspective provided by the tradition and what is true (according to the ascriber): the commitments the ascriber herself is prepared to undertake and defend. This is the form in which one engages a tradition in a dialogue aimed at deciding what commitments one ought oneself to undertake.