Logic, Language and the Question of Method in Heidegger Forthcoming in Heidegger on Logic, Casati and Dahlstrom (eds.), OUP Sacha Golob (King’s College London) This paper examines the relationship between “logic,” language, and methodology in Heidegger. I begin by contrasting two ways in which one might understand that relationship: Dummett’s position as articulated in The Logical Basis of Metaphysics and Dreyfus’ influential reconstruction of Sein und Zeit. Focusing on Sein und Zeit §33, I distinguish Heidegger’s own view from each of these. First, drawing on his discussions of “grammar,” I show where and why he diverges not just from someone like Dummett, but also from Kant. Second, I argue for the difference between my approach and the Dreyfusian one: for Dreyfus, Heidegger’s attack on logic is ultimately a question of content, for me it is ultimately a question of method. I close by indicating how this analysis might be extended to texts from the 1924 Platon: Sophistes lectures to Die Sprache in the 1950s, paying particular attention to the concept of a “metalanguage.” 1. Dreyfus, Dummett and the Philosophical Role of Language “Analytic philosophy” is a highly contested category. But historically one of its central markers has been a certain view of language. As Dummett put it: The fundamental axiom of analytical philosophy [is] that the only route to the analysis of thought goes through the analysis of language. (Dummett 1994, 128) This is not much use as a gloss on how the label is actually applied: as Dummett admits, Evans is not an analytic philosopher on his construal, but The Varieties of Reference is surely not what people have in mind when they praise or damn “continental” thinkers. Nor am I concerned in any strict sense with the causal history of “analytic philosophy” and Dummett’s place in it. Rather, I want to use Dummett as exemplifying one particular, and highly influential, strand of the analytic approach. It is a strand defined not by the details of Dummett’s own commitments, but by its basic methodology. For example, it yields the “Good Old-Fashioned Oxford Philosophy (GOOP),” as Noë recently dubbed it, of Stanley and Williamson’s intellectualism. I have referred to Stanley and Williamson as practicing GOOP. But really, what they practice is something like good old-fashioned Oxford philosophy all souped-up with contemporary linguistics. But new-fangled GOOP has many of the same old problems as old-school GOOP. The biggest problem with GOOP is that it directs our attention to considerations about language (how people talk), when theorists of mind (in philosophy or cognitive science) are interested in human nature and the nature of mind. (Noë 2005, 288) The approach exemplified by Dummett, and rejected by Noë, is thus characterized by privileging linguistic analysis as a philosophical guide. Of course, this does not mean that surface grammar can simply be taken as a reliable indicator. Rather, the method, and here one can see its history back through Russell and others, mandates an absolute focus on the relationship between such surface grammar and language’s underlying logical form. It is in this context, despite his many differences from Dummett, that we need to see events such as Carnap’s famous attack on Heidegger. As Carnap put it:  It may happen that such a sequence of words looks like a statement at first glance; in that case we call it a pseudo-statement. Our thesis, now, is that logical analysis reveals the alleged statements of metaphysics to be pseudo-statements. (Carnap 1959, 61) My interest here is in Carnap’s remark as exemplifying the “analytic” approach I am charting; I cannot address the details of his attack on Heidegger, which hangs in large part on the differences between what they understand by ‘metaphysics’.  Dummettian analytic philosophy thus simultaneously privileges language and sees a certain lack of attention to it as a founding philosophical sin: as Frege himself put it, “a great part of the work of the philosopher consists in…a struggle with language” (Frege1979, 270). The use of the formal tools and techniques of modern logic therefore becomes essential to balancing these tensions. By extension, authors such as Heidegger are natural targets precisely because they fail to employ such methods. This makes them vulnerable to accusations such as Carnap’s, that it is a flawed relationship to language that fatally undermines their thought. I now want to introduce a very different approach, set out by Dreyfus in dialogue with Heidegger, and elaborated with enormous sophistication by those influenced by him. At its core is the belief that certain levels of experience have been neglected by traditional philosophy. Dreyfus’s 2005 Presidential Address to the American Philosophical Association gives a good flavour of the key idea: But, although almost everyone now agrees that knowledge doesn't require an unshakeable foundation, many questions remain. Can we accept McDowell's Sellarsian claim that perception is conceptual “all the way out,” thereby denying the more basic perceptual capacities we seem to share with prelinguistic infants and higher animals? More generally, can philosophers successfully describe the conceptual upper floors of the edifice of knowledge while ignoring the embodied coping going on the ground floor; in effect, declaring that human experience is upper stories all the way down? This evening, I'd like to convince you that we shouldn't leave the conceptual component of our lives hanging in mid-air and suggest how philosophers who want to understand knowledge and action can profit from a phenomenological analysis of the nonconceptual embodied coping skills we share with animals and infant. (Dreyfus 2005, 49) Similarly, in his enormously influential commentary on SZ, Dreyfus talks of the need to get back to “a more basic form of intentionality” than that studied by the canon (Dreyfus 1991, 3). Dreyfus’s own stance on language is at times unclear, since it seems to fall on both sides of the “coping” framework he often uses. For example, Dreyfus draws a distinction between “the practical function of language” where it “functions as equipment” and the “thematizing use of assertions” in “theoretical reflection” (Dreyfus 1991, 208, 212). But where does, say, fully absorbed discussion of pure mathematics fit in? More broadly, Dreyfus’s account relies on a theory of breakdown or disturbance in which the “available” or zuhanden is equated with the pre-breakdown level and the “occurent” or vorhanden with explicit, “post-breakdown” awareness (Dreyfus 1991, 208). Blattner offers acute criticisms of this larger framework with which I am in agreement (Blattner 1995, 325-6). For a more detailed analysis of the problems with Dreyfus’ various formulations, see Golob 2014, 25-47. But there is a striking tendency in his work and that of others influenced by him to contrast the primitive level of “nonconceptual embodied coping skills” with linguistic content. This, for example, is how he casts his relationship with Brandom and Sellars: Phenomenologists therefore disagree with conceptualists in that phenomenologists claim that a study of expertise shows that nameable features are irrelevant to the current state of mind of the [chess grand] master when he acts…If, as Robert Brandom claims, “Sellars' principle [is] that grasping a concept is mastering the use of a word” then, according to Sellarsians, master chess play is nonconceptual. Yet clearly, what is given to the chess master in his experience of the board isn’t a bare Given…A “bare Given” and the “thinkable” are not our only alternatives. (Dreyfus 2005, 56) Sellars and Brandom are obviously very different thinkers from Dummett. But what is striking is how close the Dummettian picture of analytic philosophy is to Dreyfus’s own view of that movement: in both cases, the defining assumption is that concepts are to be analyzed in linguistic terms. Dreyfus is happy to accept that this is a legitimate enterprise; his complaint is that this focus on the “nameable” misses something more fundamental: while “analytic philosophers…continue their work on the upper stories of the edifice of knowledge, perfecting their rigorous, fascinating, and detailed accounts of the linguistic, conceptual, and inferential capacities,” they have ignored the “non-linguistic, nonconceptual discriminations” unearthed by phenomenology (Dreyfus 2005, 62). For Dreyfus, the problems exemplified by “analytic” philosophy are not, however, unique to that movement. Indeed, pre-Heideggerian phenomenology falls into much the same trap. Carman gives perhaps the most elegant formulation of this aspect of the view: Husserl’s theory of intentionality thus stands as perhaps the supreme expression of the semantic paradigm in the philosophy of mind. Unlike empiricist versions of the theory of ideas, which construe mental representations on analogy with pictures or images, the semantic model conceives of mental content in general…on analogy with linguistic meaning. (Carman 2008, 18). Heidegger, in contrast, recognised the vital need to avoid “any surreptitious reading of the structures of propositionally articulated thought back into” the explanatorily basic levels of experience (Carman 2003, 217). Avoiding this danger is vital if we are, in line with Dreyfus’s ambition, to do justice to the primary forms of intentionality: [I]intentional attitudes and experiences do not…– pace Sellars – typically contain propositional claims (Carman 2003, 217). What we have seen so far is the way in which the Dreyfusian approach defines itself by its stance on language: both “analytic philosophy,” understood in an essentially Dummettian way, and Husserlian phenomenology fail because they bought into the “semantic paradigm.” As Dreyfus always stresses, the claim is not that these schools are simply wrong: they do help delineate the “upper stories” of experience. But they simultaneously risk radically misconstruing “the ground floor” of being-in-the-world on which thought and language depend. This fundamental move has extensive implications for everything from the philosophy of action to epistemology. Consider this from Cussins: Many years ago, I used to ride a motorcycle around London. And I would often exceed the speed limit. One time a policeman stopped me and asked, “Do you know how fast you were travelling?”…On the one hand, I did know, and know very well, how fast I was travelling. I was knowingly making micro-adjustments of my speed all the time in response to changing road conditions. These micro-adjustments weren’t simply behaviours, the outputs of some unknown causal process. They were, instead, epistemically sensitive adjustments made by me, and for which I was as epistemically responsible as I was for my judgements. On the other hand, I did not know how fast I was travelling in the sense of the question intended by the policeman…the speed of my motorcycle was not made available to me as that which would render true certain propositions, and false certain others. The speed was given to me not as a truth-maker – for example, a truthmaker of the proposition that I was exceeding the speed limit – but as an element in a skilled interaction with the world. (Cussins 2003:150) For Cussins, what is needed is a philosophy of action that operates at the level of “micro-adjustments” not propositions: it is precisely this that the Dreyfusian program aims to provide. 2. Sein und Zeit on Language and Assertion: The Content Model With this sketch of the Dreyfusian approach in place, I want now to look more closely at how it treats language within early Heidegger. I will then explain both where I think that approach goes wrong, and what a better alternative might look like. In the preceding section, I stressed the explanatorily derivative status ascribed to language within the Dreyfusian model: it is dependent on the “lower floors” of non-linguistic capacities. And this clearly has some support in Heidegger’s text: most famously, Sein und Zeit §33 “Assertion as a derivative mode of interpretation.” We can press the point further: the Dreyfusian approach cashes the derivative status of language in terms of a particular story about different forms of meaning or content. Consider, for example, this from Carman’s commentary on SZ: Heidegger associates the idea of “dimming down” with assertion at SZ 156. Predicative assertions, that is, let things be seen in a specific light as this or that. Dimming down and so letting things be seen…is a kind of abstraction or decontextualisation against a background of prior practical familiarity. Propositional content therefore derives from a kind of privation, or perhaps a refinement or distillation, of practical interpretative meanings. Indeed “levelling down” the interpreted intelligibility of entities of all kinds to mere determinations of [present-at-hand] objects is “the speciality of assertion” (SZ 158). (Carman 2003, 219) Carman uses “occurrence” for Vorhandenheit and its cognates: I have altered the citation to allow terminological continuity. Likewise, this from Wrathall: In natural perception, then, we ordinarily perceive a whole context that lacks the logical structure of linguistic categories. When we apprehend things in such a way as to be able to express them in assertion, however, the act of perception is now brought under the categories of the understanding…Thus assertion manifests things differently than they are given in natural perception…This allows us to see an object with a thematic clarity that is not present in our natural perception of it. (Wrathall 2011, 20) When Heidegger talks about the derivative status of Aussage, he leaves it unclear what exactly his target is and why. Is the problem, for example, with assertion as opposed to questions or suggestions? Or is it with something much more general like propositionality? What we see in these remarks from Carman and Wrathall, and in Dreyfus’s Presidential Address, is the decision to gloss this derivative status in a particular way: there is a kind of basic meaning, the practical or perceptual, which language cannot capture. This move is vital to the structure of the Dreyfusian interpretation for two reasons. First, it allows an immediately satisfying explanation of Heidegger’s insistence that “assertion” is both derivative and that it is linked to an ontology of the present-at-hand, an ontology of “things” [Dinge]. I use Ding as it is used in texts like SZ or GA41 (see, for example, SZ 80 or GA41, 60-2). Heidegger later changes his valuation of the term radically based on its supposed Germanic heritage (GA7, 176). This dual claim is evident both in the key section of SZ, for example SZ 157-9, and in many other texts, such as GA41, 62–4 or GA29/30, 419. Within the framework defended by Carman or Wrathall, the link is straightforward: assertions reduce entities to “presence-at-hand” precisely because assertions distort or at best “distil” the irreducibly rich perceptual and practical content of being-in-the-world. This is why propositional intentionality always implies a “narrowing of content” (SZ 155). This is a significant exegetical achievement of the Dreyfusian school. After all, there is no other obvious reason why just making assertions should confine you to any particular ontology: surely, I can talk about everything from emotions to numbers to tools to Dasein? Second, the move positions Heidegger in a specific way within the canon, namely as a phenomenologist engaged in a project very much like Merleau-Ponty’s. Indeed, as Carman acutely notes: Although Heidegger is the figure to whom Dreyfus most frequently appeals, his argument frequently draws more directly from the Phenomenology of Perception than from Being and Time. (Carman 2008, 224) Bringing these points together with those of the preceding section, we can see the power and sophistication of the Dreyfusian position on language. Philosophically, it positions Heidegger as having grasped capacities which are explanatorily prior to those studied by either Husserl or analytic authors; exegetically, it seems to make excellent sense of SZ §33, including its otherwise puzzling insistence that language links to the present-at-hand. The problem, I will now argue, is that things are not quite as neat as they seem. 3. Sein und Zeit on Language and Assertion: The Methodological Model I want to begin with the link postulated by texts such as SZ 157-8 between assertion and the present-at-hand: assertion somehow pushes the world back into the “uniform plane of that which is merely present-at-hand” (SZ 157-8). In what follows, I extend some arguments sketched in Golob 2014. Elsewhere Heidegger frames the point in terms of Dinge: We cannot emphasize this fact too often: those determinations which constitute the being of the thing [i.e. katagoria] have received their name from assertion [i.e. kataphasis]…The fact that since then in Western thought the determinations of being are called “categories” is the clearest expression of the point I have already emphasized: that the structure of the thing [Ding] is connected with the structure of the assertion. (GA41, 62–4; similarly, GA29/30, 419) Earlier Heidegger defined a thing as “the present-at-hand bearer of many present-at-hand yet changeable properties” (GA41, 33), and the corresponding passage at GA25, 295–6 uses “present-at-hand” directly, so for current purposes I will take the two formulations as equivalent. The idea of a connection between certain forms of representation and certain ontologies is, of course, also central to Heidegger’ later work and its attack on Vorstellen (for example, GA5, 305). So, the claim that language might mandate a particular ontology has a comforting Heideggerian familiarity to it. But, this should not blind us to the fact that we face immediate problems when we try to get clear on what exactly SZ is saying. First, even within a single period the key terms typically lack a stable meaning. For example, “present-at-hand” and “thing” are associated variously with substances in the Aristotelian, Cartesian or Kantian senses (SZ 318; GA20, 232–3; GA41, 62–4, 107-8; GA25, 295); with entities individuated by their spatio-temporal or causal properties (GA20, 49–50; SZ 361), and with entities cut off from the network of relations that define the Heideggerian world (SZ 83–6, 157–8). But none of these are remotely equivalent. For example, substantiality is neither necessary nor sufficient for individuation by spatio-temporal or causal properties as both empiricists and Leibnizian monads can attest. Likewise, being cut off from the world is neither necessary nor sufficient for being individuated by spatio-temporal or causal properties: consider some aleph number divorced entirely from our social practice or any world which makes sense of objects precisely in terms of their primary qualities (a building yard, for example). Second, the basic idea remains puzzling: why should merely making assertions about something commit me to a particular ontology for it? Consider some of the disambiguations of “presence-at-hand” which I just flagged. Why should assertions about loneliness or number theory or that damn noise from my neighbour force me to see emotions as substances or numbers as spatio-temporally individuated or that racket as cut off from the world? The Dreyfusian approach, as we saw, explains this lacuna in terms of the rich perceptual content of lived experience; such assertions necessarily arise from a kind of “privation, or perhaps a refinement or distillation” of some richer form of meaning (Carman 2003, 19). But here is another option: perhaps Heidegger’s point is not about propositions or assertions or linguistic content at all, but rather about a particular philosophical theory of those things? As I read him, many of the issues which Heidegger raises with language are not problems with assertions or propositions per se, but rather with a particular way of thinking about them, a particular methodology. There is a genuine danger in forcing Heidegger into the idiom of analytic philosophy, but it also has some benefits of accessibility: if one were to do so here, my claim would be that his problem is not with language at all, but rather with a specific “meta-language.” I am indebted to audiences in Paris and Oxford for discussion of the idea of a meta-language in this context. Heidegger’s blanket term for the suspect philosophical approach, or the suspect “meta-language” if one is willing to talk in those terms, is “logic.” Hence his aim is to “liberate grammar from logic” by exposing the link between logic and the present-at-hand (SZ 165). I now want to show how this might help make sense of central texts such as SZ §33. I will begin with the key passage at SZ 154–5. Note first that assertion’s “primary signification,” “pointing out,” comprises not just my indicating an entity, but also includes my saying something about it: “the hammer is too heavy” (SZ 154). Since the “primary signification” of assertion says something about something, it constitutes a declarative sentence and so propositional content as that idea is standardly understood: in other words, we already have propositions in place in place several pages before Heidegger turns to the problems of SZ 157-8. The second signification, which Heidegger labels ‘‘predication,’’ (note the scare quote punctuation), then introduces “a narrowing of content as compared to the … first signification” (SZ 154–5). It is this second signification which “dims down” or “restricts” our view. In other words, contra Carman as cited above, “dimming down” is something which is done to, not by, propositional content. This occurs insofar as “logic” assumes a particular way of understanding how the assertion works: Prior to all analysis, logic has already understood “logically” that which it takes as its theme, for example “the hammer is heavy,” under the heading of the “categorical statement.” The unexplained presupposition is that the “meaning” of this sentence is to be taken as: “This thing – a hammer – has the property of heaviness.” (SZ 157) What is at issue here is a particular method for analyzing assertions, a method which generates what Heidegger calls “theoretical assertions” (SZ 157). This refers not to theory in the sense of the natural sciences – the predicate is still ‘heavy’ as opposed to ‘having mass’ – but rather to philosophical notions such as ‘‘categorical statement’’ (SZ 154). Note how Heidegger here again uses scare quotes: we can now see that their function is to flag the suspect theoretical terms, the suspect parts of what in analytic jargon would be the “meta-language.” Having introduced “logic,” Heidegger starts to flesh out its ontological implications. This finally gives us the full context for his remarks concerning “the uniform plane of that which is merely present-at-hand”: The entity which is held in our fore-having – for instance, the hammer – is initially ready-to-hand [zunächst zuhanden] as an item of equipment. If this entity becomes the “object” of an assertion [“Gegenstand” einer Aussage], then as soon as we begin with this assertion, there is already a changeover in the fore-having. The ready-to-hand entity with which we have to do or perform something, turns into something “about which” [“Worüber”] the assertion that points it out is made … Only now are we given any access to properties or the like. When an assertion has given a determinate character to something present-at-hand, it claims something about it as a “what” and this “what” is drawn from that which is present-at-hand as such. The as-structure of interpretation has undergone a modification. In its function of appropriating what is understood, the “as” no longer reaches out into a totality of involvements. As regards its possibilities for articulating reference-relations [Verweisungsbezügen], it has been cut off [abgeschnitten] from that significance which, as such, constitutes environmentality. The “as” gets pushed back [zurückgedrängt] into the uniform plane of that which is merely present-at-hand. It sinks to the structure of just letting one see what is present-at-hand in a determinate way. (SZ 157–8, original emphasis) We can immediately note the scare quotes again picking out the key terms of the problematic approach; he also uses italics here for the same purpose, for example with ‘properties.’ Heidegger’s claim here is that giving philosophical weight to this type of theoretical framework not only alters the way in which assertions are understood but equally modifies the way in which one views the entities intended by those assertions: they are thematized as ‘‘objects’’ of ‘‘categorical statements’’ and bearers of “properties.” How exactly does this work? A full discussion would require close treatment of both the ambiguities in Heidegger’s talk of the “present-at-hand” and his views on modern symbolic logic, which he knew partly through Cassirer. But I can indicate the basic point by using two examples. First, consider the conception of the present-at-hand as a substance in an Aristotelian, Cartesian or Kantian sense. Heidegger is pointing out the close links between the ontologies of these thinkers and a philosophy of language based around a combinatorial analysis of propositional form that privileges the subject–predicate structure. The classic example is Kant’s claim in the Metaphysical Deduction to have derived the categories from the logical forms of judgments. Kant 1902, KrV A70–80/B95–106. Heidegger sees exactly the same problem in Leibniz: Leibniz sees that this interpretation of substance takes its bearings from predication and therefore a radical determination of the nature of predication, of judgment, must provide a primordial conception of substance…Here the ontic subject, the substance, is understood from the viewpoint of the logical subject, the subject of a statement. (GA26, 41–2) He views others, such as Aristotle, as making the same move albeit in a less systematic fashion (GA41, 62–4; GA25, 295). These authors happily endorse the link between assertion, “logic,” and the present-at-hand because they regard the resultant substance ontology as a positive: what Heidegger is doing is using that very same ontology as the basis for a modus tollens against the underlying philosophy of language. In making this move, Heidegger is in good company: thinkers from Nietzsche to Russell have made similar observations. As Russell puts it, in perhaps the sole line in his corpus that could equally have been written by Heidegger: The ground for assuming substances – and this is a very important point – is purely and solely logical. (Russell 1937, 49) By extension, for Nietzsche and for Russell, classical subject-predicate logic is both misleading and ontologically dangerous. On Nietzsche, see Nietzsche 2002, §16 and Twilight of the Idols in Nietzsche 2005, III/5. This is exactly the point Heideggerian is making. This brings me to the second case: how should we understand Heidegger’s attacks on “logic” once we move outside the pre-Fregean subject-predicate framework? Heidegger is consistently scathing of modern logic, so it is clear at least that he does not think, as Russell did, that modern logicians have fixed the problem (SZ 88). Matters here are complex and a full treatment would require a detailed discussion of Heidegger’s broader philosophical methodology. I have tried to supply the necessary context in Golob 2014, 50-62. But the basics are clear enough: Heidegger is worried that a focus on the logical structure of the assertion will necessarily lead to a concomitant neglect of the existential, i.e. the pragmatic social and environmental context in which in which language actually functions. It is in this sense that entities within a logic-dominated philosophy are “cut off” from their “significance” and “pushed back into the uniform plane of that which is merely present-at-hand” (SZ 158). An analysis such as Russell’s would no longer talk of a “categorical statement”; as Russell stressed, such outdated logic is utterly unable to deal with the inferential status of relational properties. Heidegger’s worry, however, is that the modern ingenuity expended on analyzing such properties, on grasping, the “empty formal idea of relation,” ironically leads to the “suppression of the dimension within which the relevant relation can be what it is,” namely, its worldly context (GA29/30, 424). By extension, his own preferred alternative is to approach assertion from within and only from within an analysis of Dasein – to start not with syntactic or semantic form, but with the various roles of assertion within the rich social and instrumental context in which Dasein deploys it. Thus, we are far better off with Aristotle’s Rhetoric than we would be if he had written a philosophy of language since the former treats speech “as a basic mode of the being of the being-with-one-another” (GA18, 171). As he puts it himself, his aim is to move “from the question of what language is to the question of what man is” (GA38, 38). If we now turn back to the Dreyfusian account, we can see why my position is strikingly different. In Dreyfus the link between assertions, propositions and the present-at-hand was explained in terms of differing forms of content or meaning, specifically the inability of propositional intentionality, and by extension assertion, to capture perceptual or practical awareness. On my approach, in contrast, things are very different. Propositions and assertions are not the problem, and there is no story about inexpressible meaning. The problem is rather with a particular way of thinking about language. It is an issue of method. The distinction between the two approaches is real. First, my view avoids trapping Heidegger in a self-reference paradox: he refers to his own claims as both “propositions” and “assertions,” a fact which would, on the Dreyfusian approach, imply that merely in writing about Dasein he had already levelled it off (GA24, 461). Second, I avoid the pressure created by the Dreyfusian approach to conflate Heidegger and Merleau-Ponty: after all, the perfect candidate for a primal form of intentionality uncapturable by language is the fine-grained meaning characteristic of perceptual motor-intentionality. Such a conflation is characteristic of the Dreyfusian reading and it is a mistake. As I put it elsewhere: The problem is that SZ would, effectively, state Heidegger’s views without giving any argument for them: there is little discussion there of the type of detailed motor intentional case study needed to motivate the view that such content even exists. This absence is even more striking in Heidegger’s other works. For example, GA3 and GA25 detail his disagreements with Kant, yet the body and embodiment receives no treatment at all. Could one conceivably have said the same of a similarly extensive confrontation with Kant written by Merleau-Ponty or Todes?... Ultimately, an appeal to motor intentionality risks turning…Heidegger’s key arguments into… a promissory note to be cashed by the Phenomenology of Perception. (Golob 2014, 46) On my account, in contrast, Gadamer is a much more natural interlocutor: the task in texts such as SZ §33 is primarily “to liberate the verbal nature of understanding from the presupposition of philosophy of language” (Gadamer 2006, 404). I deliberately say only “interlocutor”; I certainly do not think that Heidegger shares all of Gadamer’s positive views on language. As Dreyfus presents it in his Presidential Address, analytic philosophy has done a fine job of elaborating the “upper floors” of experience, advancing “rigorous, fascinating, and detailed accounts of the linguistic, conceptual, and inferential capacities that are uniquely human”: the problem is that they have missed the “lower stories.” But this is very far from Heidegger’s view. The problem as he sees it is not that we have a decent philosophy of language but have missed something else. It is rather that the existing philosophies of language amount to “a monstrous violation of what language accomplishes” (G36/37, 104). The distinction between my reading and the dominant Dreyfusian one is thus a genuine one. But it can also be easily obscured. This is because there is a natural tendency to talk as if the suspect methodology were automatically built into terms such as ‘proposition.’ Using this idiom, one might say that my account, like the Dreyfusian one, also regards propositions as necessarily derivative and necessarily tied to presence-at-hand. Heidegger himself is deeply conflicted on this and on the general question of existing philosophical terminology. This is exemplified in his treatment of terms such as logos which he alternately tries to retrieve for his own purposes and to replace with coinages such as Rede (compare GA40, 128 and 194). However, I think it much better to separate out propositions per se, i.e. the content of declarative statements, from the pernicious theory which has accompanied them. Most obviously, failing to do so obscures the points just made about the differences between my approach and the Dreyfusian one. Such a failure also makes dialogue with analytic philosophy extremely hard. This is because, whatever one’s view on the standard Fregean/Russellian/Evansian/possible worlds options, almost all analytic authors will regard “the noise downstairs is very loud” as a proposition – and this kind of claim was never the target of SZ §33. While the Searle-Dreyfus dialogue was complicated by many issues, one of the most persistent was a simple misunderstanding as to what each author meant by basic terms like ‘proposition’: we should avoid framing things in a way that will systematically generate such misunderstandings (see, for example, Searle 2000). 4. Extending the Methodological Approach to Heidegger’s Later Work In line with the scope of this paper, I have focused on SZ. I will now indicate how my reading might be expanded to Heidegger’s later work. Heidegger’s “later work” is, of course, a blanket way of describing a vast range of stylistic, terminological and conceptual shifts, to which I cannot do justice here. But, one can see how the argument I picked out reoccurs throughout Heidegger’s corpus: indeed, it is, in my view, one of the few absolute points of continuity. For example, three years before SZ, he rails against the way Satzlogik has distorted language: As orientated in this way, i.e. as taking the theoretical proposition for its exemplary foundation, propositional logic [Satzlogik] at the same time guided all reflections directed at the explication of logos in the broader sense, as language [Sprache], and insofar as it did so the whole of the science of language, as well as, more generally, the entire philosophy of language, took their orientation from this propositional logic. All our grammatical categories and even all of contemporary scientific grammar – linguistic research into the Indo-Germanic languages etc. – are essentially determined by this theoretical logic. Yet there does indeed exist the task of conceiving logic, once and for all, much more radically than the Greeks succeeded in doing and of working out thereby, in the same way, a more radical understanding of language itself and consequently also of the science of language. (GA19, 253) This is exactly my claim: the key contrast is between language and pernicious theories of it. The same point is also clearly visible in texts from the mid 1930s. I mentioned above the 1933/34 lectures Sein und Wahrheit (GA36/37). The problem, Heidegger states, is that “grammar” was dominated by “logic”: the result is a very specific “representation of language,” one that amounts “to a monstrous violation of what language accomplishes” (GA36/37, 104). In place of this theory, Heidegger enjoins us to “consider a poem or a living conversation between human beings” (G36/37, 104; similarly, GA54, 102). This is the strategy I suggested Heidegger had in mind in SZ §33, a philosophy of language grounded on our social practices of assertion and speech. The task, in short, is “destabilizing the grammatical representation of language” (GA36/37, 104). Another particularly interesting case is the lecture course GA38, Logik als die Frage nach dem Wesen der Sprache, delivered in the Summer of 1934. This illustrates the complex development of Heidegger’s thought during the mid 1930s: recognizable themes from SZ, such as the Augenblick or Selbstverlorenheit are interwoven with new vocabulary, in particular that of the Volk (GA38, 50, 57). It is, rightly, hard to read Heidegger’s casual examples of SA-Dienst, but for current purposes what is important is the absolute consistency of Heidegger’s position on the status of language. The task, as he sees it, is to revive “logic,” as the study of logos, and to “ask about the essence of language.” One sees again his ambivalent stance on existing terminology: in this text, even “logic,” rightly construed, can be saved. The difficulty, precisely as above, is that language has been radically misconstrued. Hence, we ask first of all about the essence of language [Wesen der Sprache], but not via a philosophy of language, which degrades language to a specific, separate area. (GA38, 30) Only by doing this can we understand the nature of the human being: indeed, a grasp of that and a genuine grasp of language are inseparable (GA38, 31). Exactly as in SZ, the problem is thus not language as opposed to perception, but language rightly understood as opposed to language as conceived by the tradition. And, exactly as in SZ, only by understanding Dasein and language together can progress be made. We see exactly the same points again in Die Sprache from 1950. The orthodoxy that Heidegger rejects is one on which: [A]ll statements are referred in advance to the traditionally standard way in which language appears. The already fixed view of the whole nature of language is thus consolidated. (GA12, 13) How far can we push these claims of continuity once we get to Heidegger’s post-war period? One feature of my account is that it becomes possible to understand many of his various writings as more closely continuous than is often believed. After all, if SZ’s warnings really are directed against language per se, as on Dreyfus’ model, it is harder to explain its later role as the “house of being” without positing a radical break. In contrast, on my approach one could read even SZ as accepting the same foundational role for language visible in the later works – it would oppose only the misguided methodology sketched above. I do not however, want to make that further claim here, in part because I agree with Wrathall that the meaning of terms such as Sprache in the later work is complex; in part, because that would require a vastly more detailed treatment of SZ. But what we can say is this: a central concern in Heidegger’s thought from the 1920s through the post-war period was to identify a series of problems not with language but with the ways in which language had been conceived, systematized, and misconstrued. In SZ, these worries are framed in terms of the present-at-hand; in the latter work, they are often put in terms of the threat of technology. In both cases, they are often summarized by talk of “logic.” One can also see how deep the worries run. For example, as noted, one natural way to articulate Heidegger’s point in an analytic context is by talk of a “meta-language.” The problem from a Heideggerian perspective is that such philosophical idioms are never simply neutral or transparent; indeed, he sees that very phrase as itself bound up with the underlying problems he traces, with an attitude that misconstrues language and – by extension – the entities which it names. As he characteristically puts it, “meta-language and Sputnik are… one and the same” (GA12, 160). Later Heidegger thus seeks both a new understanding of language and a new way of articulating that understanding within a transformed philosophy. It is in this sense that: The liberation of language from grammar into a more original essential framework is reserved for thought and poetic creation. (GA9, 314) We see here both the difficulty Heidegger faces in articulating his views and the ease with which he slides between different formulations. SZ sought to “liberate grammar from logic” (SZ 165): the very same task is now presented as liberating “language from grammar”, that is from a conception of grammar corrupted by traditional philosophy. I am indebted to Daniel Dahlstrom for highlighting this point. 5. The Apophantic and the Hermeneutic “As” I want to close with one final issue: the status of the “as” – what Derrida rightly called the “great phenomenological-ontological question” (Derrida 1992, 289). How does my position bear on this? As we will see, this is intricately related to another question: what would happen if Dreyfusians were to simply adopt my gloss on SZ §33? Consider, for example, the distinction SZ 158-9 draws between the hermeneutic and the apophantic “as”. One simple option would be to use my approach and then to align the “hermeneutic” with assertions properly understood, and the “apophantic” with those treated in terms of the problematic philosophy. But I think things are more complex – Heidegger uses “apophantic” and its cognates in SZ 154-5 to introduce both the basic features of assertion and its gradual appropriation by the problematic theory. Instead, as I see it, the reference to the “hermeneutic as” alludes to a further claim, namely, that assertion, even when correctly understood, still remains derivative on some prior form of intentionality. It is this further claim which is in play in passages such as the following: The proposition “a is b” would not be possible with respect to what it means, and the way in which it means what it does, if it could not emerge from an underlying experience of “a as b.” (GA29/30, 436) Compare GA26, 158. Now one might wonder whether in accepting this I have suddenly acquiesced in the Dreyfusian interpretation? After all, that interpretation also viewed language as derivative. The answer is no. On my account, there are three things in play and it is vital to keep them distinct: (1) the “as” structure, identified by Heidegger as the primary level of intentionality, (2) assertion, and (3) assertion as construed by the philosophically problematic methodology. Heidegger holds that (2) is explanatorily dependent on (1). He further holds that (3), a distortion of (2), introduces a deeply pernicious ontology: this is the claim which links assertion and presence-at-hand. The Dreyfusian model, in contrast to my account, conflates the problems of derivation and the pernicious ontology: assertion is seen as the source of a present-at-hand framework because it is seen as derivative, unable to capture the underlying form of intentionality. Why does this fine difference in argument structure matter so much? Well, I have argued above that we should not go back to anything like the standard Dreyfusian story; we cannot, for example, impose Merleau-Ponty’s concepts onto SZ. So, whatever my story regarding (1) and (2) is, it will need to differ from that. To put the point another way, if not all assertions are linked to the present-at-hand, then the priority of (1) over (2) cannot be a function of the supposed fact that the latter is a privative representation of the former: it cannot simply be that “assertion manifests things differently than they are given in natural perception” (Wrathall 2011, 20). So, the task becomes offering an analysis of the “as”/”is” distinction in early Heidegger that recognizes both that language is not the primary form of intentionality and that language, properly understood, can nevertheless do justice to that experience. To put the point another way, we need to show how Heidegger can avoid collapsing into a McDowellian style view on which the “as” is tacitly propositional. Doing that, I have argued elsewhere, requires us to fundamentally rethink the relationship between the nonconceptual, the conceptual and the propositional. Golob 2014. My thanks to the Editors and to audiences in Paris, Oxford and Kennebunkport for their extremely helpful comments on earlier versions of this material. 11