Skip to navigation – Site map

HomeIssuesXIV-2Symposia. Pragmatism and Phenomen...Redefining the Synthetic a Priori

Symposia. Pragmatism and Phenomenology. A Metaphilosophical Reconsideration

Redefining the Synthetic a Priori

Sellars’ Via Media Between Pragmatism, Phenomenology, and Logical Empiricism
Claudia Cavaliere


At B19, Kant summarizes the general problem of pure reason in the problem of synthetic a priori judgments. The vicissitudes that have affected last century’s philosophy are, in this sense, a confirmation of its significance: the problem of synthetic a priori knowledge has indeed crossed all the major philosophical currents of the twentieth century, being treated in a wide variety of ways by phenomenologists, logical empiricists, and pragmatists. One of the most original treatments of the issue is that offered by the young Wilfrid Sellars between the late 1940s and the early 1950s, consisting of a hybridization of different traditions. In this paper I will present Sellars’ unique re-elaboration of the notion of the synthetic a priori as it is elaborated in his early writings, showing how his interweaving of themes from pragmatism, phenomenology and logical empiricism leads not just to what I believe is the only possible approach for a realist philosopher, but also to an unexpected Kantian solution in spirit.

Top of page

Full text

  • 1 Citations to Sellars’ works will use the now standard abbreviations as listed in the References. Fo (...)

1If we are to take seriously what Kant says at KrV B19, then we agree with the fact that the problem of Pure Reason is the problem of synthetic a priori judgements.1 The sixth section of the B Introduction declares indeed:

One has already gained a great deal if one can bring a multitude of investigations under the formula of a single problem. For one thereby not only lightens one's own task, by determining it precisely, but also the judgment of anyone else who wants to examine whether we have satisfied our plan or not. The real problem of pure reason is now contained in the question: How are synthetic judgments a priori possible? (KrV B19)

2By giving a quick look to the history of the last two centuries’ epistemology, the central role of this question is confirmed by the fact that, as Coffa puts it, every philosophical development since 1800 was in form of a reply to Kant and specifically to the problem of synthetic a priori knowledge (Coffa 1991: 7). After the Einsteinian revolution, the problem became even more urgent: once the structure grounding the Kantian system was dismissed and the possibility of multiple alternative logical systems was envisaged by the discovery of non-Euclidean geometries, many philosophers thought the problem of knowledge could be resolved in an easier way by simply getting rid of the notion of synthetic a priori, arguing for the sufficiency of analytic a priori and synthetic a posteriori knowledge.

  • 2 Schlick 1949.
  • 3 It goes without saying that not all logical positivists shared the same attitude towards the distin (...)

3The most famous controversy on the matter involved the two greatest European philosophical movements of their time: namely, phenomenology and logical empiricism. The quarrel started in 1931 with the publication of Moritz Schlick’s influential article “Gibt es ein Materiales Apriori?”2 Schlick targeted the Husserlian notion of material a priori and, more broadly, the phenomenological approach to the analytic/synthetic distinction altogether. Whereas phenomenologists insisted in talking about synthetic a priori propositions, the fundamental claim of the new version of empiricism Schlick was espousing was precisely that there can be no such propositions.3 Moreover, according to Schlick, since the term a priori was tied to concepts such as “analytic” and “formal,” the concept of a synthetic a priori – which involved the idea of a form of necessity whose source is extra-logical – amounted to a contradictio in terminis.

4If the division of knowledge between analytic a priori and synthetic a posteriori advocated by empiricists may have looked like the easiest solution to many, among the ones who were convinced it was possible to have their own cake and eat it too was Wilfrid Sellars. Sellars was indeed convinced that the empiricists’ rejection of any kind of necessity that wasn’t strictly logical or formal was too quick, and ultimately couldn’t account for the distinctive necessity characterizing laws of nature. On the other side, he agreed with the empiricists that classic “rationalist” conceptions of extra-logical necessity had by then become so foggy that, if we were to save at least “the fire burning under the smoke” (LRB: §39), a radical change was needed. As Sellars states in one of his earliest publications, “rationalism gave the grammar, but contaminated it with platonizing factualism. Classical empiricism threw out the platonizing, but continued to factualize, and confused the grammar of philosophical predicates by attempting to identify them with psychological predicates” (RNWW: §4). If deflating the platonizing lexicon was a key move required to open the way for a fully naturalistic ontology – the irrevocable starting point of Sellars’ philosophical reflections – the empiricists’ attempts ultimately went too far in throwing out the baby with the bathwater. What was needed for a renewed and more effective, genuinely philosophical empiricism was to “absorb the insights of rationalism” (LRB: §19) translating them into the vocabulary made available by the new way of words Sellars was developing at the time. The notion of synthetic a priori, I will argue, was amongst the insights he intended to preserve. Sellars’ own interpretation of the notion did justice to both phenomenological claims about extra-logical necessity and the logical empiricists’ deflationary strategy, while at the same time incorporating also some core insights from pragmatism.

  • 4 The majority of Sellars’ so-called “early writings” have been collected by Jeffrey Sicha in the vol (...)
  • 5 The only close examination available for the early writings is Peter Olen’s Wilfrid Sellars and the (...)

5The texts I will consider here span roughly from 1948 to 1954 and are part of the so called Sellars’ “early” phase.4 These “early” articles are notably the most underexplored part of Sellars’ thought: the infamous aura that surrounds them – that of obscure writings among an already notoriously cryptic corpus, mostly devoted to small debates inaccessible for those without a knowledge of the historical period in which they were produced – surely contributed to their lack of fortune.5 However, besides illuminating a neglected part of Sellars’ philosophical system, my paper aims to show how some key questions of Sellars’ later philosophy – the ones that would make him the famous philosopher we know – were already simmering in the early phase.

6The itinerary I am about to outline includes several steps. First, I will provide the background against which the originality of Sellars’ discussion of the synthetic a priori can be adequately appreciated, namely: the Schlick-Husserl debate on the issue (1). Then, I will present Sellars’ own treatment of the problem of extra-logical necessity. In this context, I will focus on two key notions, namely material rules of inference and real connections. The former is introduced in the context of a linguistic analysis, precisely with the aim of addressing the issue of extra-logical necessity; the latter is used in the context of a modal analysis of the special kind of necessity underlying laws of nature (2). I will thus conclude with the last ingredient to be added to make this revised theory of synthetic a priori knowledge fit into a naturalistic framework: namely, the Lewisian pragmatic conception of the a priori, that Sellars embraces with substantial modifications (3).

1. Prequel: The Schlick-Husserl Debate on the Material A Priori

  • 6 “The empiricism which I represent believes itself to be clear on the point that, as a matter of pri (...)

7In the Introduction to the volume Readings in Philosophical Analysis – an anthology of texts from the rising “analytic” tradition – Herbert Feigl pinpointed the empiricists’ trademark by declaring that “[a]ll forms of empiricism agree in repudiating the existence of synthetic a priori knowledge” (Feigl & Sellars 1949: 13-4). This statement was explicitly borrowed from Schlick’s influential article “Is There a Factual A Priori?,”6 whose first English translation appeared within the same collection.

  • 7 To be sure, although Schlick’s article has been most often considered as a direct attack to Husserl (...)

8Schlick’s article, which thanks to Sellars’ translation would soon become a manifesto of a new generation of philosophers, was a direct attack to the use of the analytic/synthetic distinction made by phenomenologists, the most problematic instance of which was Husserl’s notion of material a priori.7 Schlick claimed the phenomenologists show a “confusion” in using the term a priori (Schlick 1949: 277) – a confusion which, in its most pernicious shortcoming, led to the formulation of “propositions of absolute validity which nevertheless have something to say concerning the stuff or material of experience” (ibid.). To clarify the meanings of the term a priori once and for all, Schlick puts forward a straightforward solution:

An analytic proposition is one which is true by virtue of its form alone. Whoever has grasped the meaning of a tautology, has in doing so seen it to be true. It is because of this that it is a priori. In the case of a synthetic proposition, on the other hand, one must first understand its meaning, and afterwards determine whether it is true or false. It is because of this that it is a posteriori. (Ibid.: 278-9)

9As the quote shows, Schlick establishes a sharp dichotomy by making the meaning of a priori co-extensive with the meaning of analytic and formal, on the one hand, and the meaning of a posteriori coextensive with synthetic and factual: in this way, all propositions are either analytic a priori (tautologies) or synthetic a posteriori, while a third option – that of synthetic a priori propositions – amounts to a plain “logical impossibility” (ibid.: 281). Therefore, the answer to the question he puts in the title of his article is a firm no: since a priori means literally “independent from experience” and the synthetic aspect of knowledge is tied to “the stuff or material of experience” (ibid.: 277), the two predicates simply contradict one another.

10The dispute between Schlick and the phenomenologists received a good amount of attention. Several scholars have pointed out how Schlick’s critique did not quite hit the target, mainly by oversimplifying a distinction that, ultimately, had its own right (Bordini 2011; Miraglia 2006; Piana 1971). In order to settle the dispute, one should first get a clear grasp of how the problem of the material a priori was originally formulated by Husserl.

11The analytic/synthetic distinction is discussed by Husserl in the Third Logical Investigation. There, Husserl gives a definition of analytically necessary propositions as

propositions whose truth is completely independent of the peculiar content of their objects […]. They are propositions which permit of a complete “formalization” and can be regarded as special cases or empirical applications of the formal, analytic laws whose validity appears in such formalization. (Husserl 2001: 21)

  • 8 That they are not logical necessities can be appreciated by noticing that their terms can’t be subs (...)

12Ironically, the definition of analyticity Husserl gives in this passage matches the one provided by Schlick: both philosophers define analytic propositions as propositions whose validity is based solely on their logical structure (“true by virtue of their form alone”). The core of the disagreement between the two lies elsewhere, namely in Husserl’s acknowledgment of another kind of propositions which, although endowed with something like the necessity of logical laws, do not fit into the definition of analyticity just given. He is thinking about propositions like “every color is extended in space,” “every sound has a pitch and an intensity,” or “one and the same surface cannot be simultaneously red and green”: these claims are clearly not obtained through induction, however, they cannot be considered as logical necessities either.8 They are, in all respects, synthetic a priori propositions, which spell out necessary features concerning the structure of our experience while yet being not analytic in the sense defined.

13Now, according to Schlick, propositions like the ones listed by Husserl only seem to give expression to material content – the impression being validated by their reference to colors and sounds – whereas actually they are nothing but tautologies. Evidence for this claim lies in their actual usage in everyday talk: it is not a coincidence that these propositions are never stated explicitly outside philosophical discourse. The reason for Schlick is precisely that they are trivial truths that say nothing about the world. In other words, according to Schlick the validity of materially a priori propositions is a logical one: to deny their truth means to violate “the logical rules which underlie our employment of color-words” (Schlick 1949: 284), in the same sense in which “if I hear that the dress was both green and red, I am unable to give a meaning to this combination of words; I just do not know what it is supposed to mean” (ibid.). He can therefore claim that:

The error committed by the proponents of the factual a priori can be understood as arising from the fact that it was not clearly realized that such concepts as those of the colors have a formal structure just as do numbers or spatial concepts, and that this structure determines their meaning without remainder. (Ibid.: 285)

14Is Schlick’s account of materially a priori propositions as “true in virtue of their form alone” overall convincing? It has not gone unnoticed that the consistency of Schlick’s argument lies entirely on his definition of analyticity (Piana 1971; Bordini 2011). In this sense, “the real problem comes down to the acceptance or rejection of their [materially a priori propositions’] analyticity” (Piana 1971: 9).

  • 9 “Form (in the sense relevant to the debate) was, for both Schlick and Husserl, conceptual or logica (...)

15If we look carefully at Husserl’s argument, it is possible to see that he was drawing attention on something that Schlick’s clear-cut dichotomy simply eludes: namely, the difference between necessity according to logical form and necessity according to material content9 or, with a more proper terminology, a distinction between formal a priori (which gives rise to analytic a priori propositions) and material a priori (which gives rise to the synthetic a priori).

  • 10 For a more detailed comment on the matter see De Santis (2015: 172ff.).

16Leaving the extremely sophisticated details of Husserl’s position aside, the kern is that whereas Schlick assumed concepts like colors or sounds to be formal concepts, so that propositions like those mentioned by phenomenologists would do nothing but express, tautologically, their grammatical structure, Husserl saw them as material concepts (sachhaltige Begriffe). The material character of synthetic a priori propositions is thus grounded in the latter’s specificity – or, in other words, it is expression of something embedded in colors and sounds themselves:10

The necessities or laws which serve to define given types of non-independent contents rest, as we often have emphasized, on the specific essence of the contents [in der Besonderheit der Inhalte], on their peculiar nature [Eigenart]. […] To these essences correspond the concepts or propositions which have content, which we sharply distinguish from purely formal concepts and propositions, which lack all “matter” or “content.” […] This cardinal division between the “formal” and the “material” spheres of Essence gives us the true distinction between the analytically a priori and the synthetically a priori disciplines (or laws and necessities). (Husserl 2001: 19)

17Husserl’s reflections on the notion of “material a priori” are indeed part of a wider, systematic framework that aimed at lying the foundation of real (or material) ontologies. Materially a priori propositions directly call for a Geseztmäßigkeit in sensible contents themselves: the impossibility to experience sounds without pitch or colors without extension does not simply stem, as Schlick maintains, from logical non-sense, but is grounded on regional ontologies which describe specific portions of reality. In this sense, “to argue for the existence of material a priori equates indeed with the thesis according to which a certain complexity belongs to the domain of (sensuous) data themselves” (Miraglia 2006: 111; my transl.). The Schlick-Husserl debate, therefore, sees two opposites stances in play:

[w]hereas Husserl’s eidetic analyses remained grounded in the examination of the specific character of particular perceptual domains and regional ontologies, Schlick’s spare and nominalist conception of logic […] eschewed the specific description of experience, preferring to operate on the level of language and understanding conceptual analysis essentially as grammatical analysis. (Livingston 2002: 264)

18In what follows, I will show how Sellars carries out his reflections on the matter operating precisely between these two poles. By trying to do justice to both the phenomenological claim for extra-logical necessity and the logical empiricist’s deflationary attitude, he crafted a notion of the synthetic a priori that draws something from both views. Before going on, though, it will be useful to add a biographical note.

2. Sellars on Extra-logical Necessity: Material Rules of Inference and Real Connections

  • 11 The recognition of Sellars’ philosophy centrality and his admission among the “canonical” philosoph (...)

19That Wilfrid Sellars has been a quite sui generis analytic philosopher is acknowledged. Having received a wide-ranging education which permitted him to effectively navigate both Western philosophy’s classics and contemporary analytic debates on logic and epistemology, he was able to develop what is now considered to be an original and insightful philosophical position.11 To make his position quite unique in an increasingly polarized philosophical scene contributed his characteristic strive towards a reconciliation between classic themes of European “continental” thought with the new analytic philosophy spread overseas. This reconciliation often took the shape of a distinctive way of relating to the history of philosophy, in which, as he would famously later say, “[t]he history of philosophy is the lingua franca which makes communication between philosophers, at least of different points of view, possible” (SM: 1).

20The profound re-elaboration behind the notion of the synthetic a priori, besides being an illustration of both his sensibility to the history of philosophy and his philosophical brilliance, is also a perfect example of his “mediating” capacity to merge insights from different traditions: in this case, phenomenology, logical empiricism, and pragmatism. If reference to pragmatism and logical empiricism probably came as natural for a North American philosopher writing in the late 40s, the real question to be asked is instead: how could Sellars have gotten in touch with phenomenology? Far from being a marginal question, the answer will provide an important piece of the general picture I am sketching.

21Recently, considerable efforts have been dedicated to the exploration of the history and reception of phenomenology in the US (Ferri 2019; Manca & Nunziante 2020). A point to be hold steady is that the whole process was everything but neutral: “in the transition from one side of the Atlantic to the other, connotations and uses of the term ‘phenomenology’ change in meaning and the very purposes of phenomenological agenda are redefined” (Nunziante 2018: 152; my transl.).

  • 12 For a deeper analysis of the relationship between Sellars and phenomenology, see Nunziante 2018 and (...)

22Now, Wilfrid Sellars’ education took place precisely during the time span in which this shift was happening. An important step in his education that will be emphatically recalled in his later Autobiographical Reflections (AR) was indeed the period spent at the University of Buffalo as Marvin Farber’s teaching assistant starting from 1933. At the time, Farber was leading figure of the phenomenology recently landed in the United States,12 having been taught himself by Husserl in Freiburg and Heidelberg during his doctoral studies in the 1920s. With Sellars’ words,

Marvin Farber led me through my first careful reading of the Critique of Pure Reason and introduced me to Husserl. His combination of utter respect for the structure of Husserls thought with the equally firm conviction that this structure could be given a naturalistic interpretation was undoubtedly a key influence on my own subsequent philosophical strategy. (AR: 75)

23Far from being a passive recipient, Farber had quite an original standpoint towards Husserlian phenomenology: the aspect worth to be highlighted – which Sellars himself recalls as a “key influence” on his subsequent philosophical strategy – was Farber’s commitment to the project of a naturalized phenomenology, which mirrored the then two indispensable prerequisites to the American philosophical discourse, namely realism and the idea of philosophy as a scientific enterprise (Nunziante 2018: 153). Farber indeed considered phenomenology more fruitful when taken as a method rather than as an autonomous discipline, and he was always critical of the more idealistic and subjectivistic aspects of Husserl’s philosophy: it was thus no accident that he had a clear preference for texts like the Logical Investigations or Experience and Judgement, while he didn’t particularly enjoy the more idealistically inclined Ideen.

  • 13 Even though this remains valid, it is also true that in the later production there are a few places (...)

24Bearing in mind Farber’s approach to Husserl is important insofar as it gives us a hint as to how Sellars relates to phenomenology himself: Farber’s conviction about the need to naturalize phenomenology in order to make it resonate with a fully scientific conceptual framework will be indeed inherited by the philosopher from Pittsburgh. In line with the approach to the history of philosophy I mentioned before, however, Sellars never commits himself to a precise phenomenological “canon.” One would search in vain for accurate analysis or discussions of text passages from one or another Husserl’s writings: what he does is, more generally, to appropriate a certain philosophical strategy that he exploits to solve his own problems.13 As he will later recall, “sheer phenomenology […] takes us part of the way, but finally lets us down” (SRPC: 178) – a clear sign that Sellars shared Farber’s idea of phenomenology as a method that would have to be supplemented with other conceptual and scientific resources according to his needs.

25In the preceding section I have sketched the main elements of the Schlick-Husserl streit. Whereas Husserl’s goal was to distinguish between two different sources of necessity – a logical one, underlying analytic a priori propositions, and a material one, underlying synthetic a priori propositions – Schlick argued that the latter only gave the impression of saying something about the world. On the contrary, however, it expressed nothing but linguistic rules governing concepts of sound or color. In accordance with the linguistic turn Sellars embraced at the beginning of his philosophical career, reflections on synthetic a priori knowledge will first grow out of his philosophy of language too. The key concept through which it is possible to read Sellars as maintaining a doctrine of synthetic a priori knowledge, I argue, is that of material rules of inference.

26To explore Sellars’ notion of material rules of inference it is necessary to go back to his earlier writings: namely, “Pure Pragmatics and Epistemology” (PPE), “Epistemology and the New Way of Words” (ENWW), and “Realism and the New Way of Words” (RNWW), in which Sellars spoke of them under the name of conformation rules. Conformation rules were part and parcel of Sellars’ sophisticated project of a pure pragmatics which consisted of, in extremely general terms, a re-formulation of basic epistemic predicates (such as “true, false, designates (or means), verifiable, confirmable, verified, confirmed, and meaningful,” PPE: §4) as meta-linguistic devices specifying “the use of any language in which empirical description is possible” (Brandom 2015: 8).

  • 14 Part of Olen (2016)’s second chapter (esp. ibid.: 23-35) is devoted to explaining how the belief th (...)
  • 15 This quote must be approached with great care. The relations of “matters of facts” Sellars is here (...)

27The main reason pure pragmatics was developed in the first place was Sellars’ belief that Carnapian semantics lacked the resources to properly tie language to the world (or, more precisely, to enable extra-linguistic reference).14 He thus collected several tools which he believed Carnap’s account needed to be supplemented with and proceeded to describe them, with slight adjustments, in the three aforementioned essays. Conformation rules, in this sense, are a pragmatic addition to formation and transformation rules: while formation rules specify the way in which symbols of a language can be united into new expressions and transformation rules concern conditions on inference, conformation rules have the particular role of imposing restrictions on the possible combinations of predicates with individual constants. Having been introduced precisely to mimic the restrictions present in our use of natural language, over and above merely formal logical constraints, conformation rules thus “specify for each non-relational predicate in the calculus, the relational predicates which can participate in sentences with one and the same individual constant which is conjoined in a sentence with the non-relational predicate in question” (PPE: §22). The thing here is to realize that, when we are talking about conformation rules, we are really talking about material restrictions – a remark which resonates with what Sellars says immediately after this, namely that they are to be distinguished from mere generalizations, “even though they are not, in the usual syntactical sense, analytic(ibid). Conformation rules, which differentiate predicates by specifying their combining properties, are indeed ultimately defined as, “to use Hume’s phrase, ‘relations of matter of fact’ in the world to which the language applies” (RNWW: §33n9).15

28The material aspect underlying conformation rules will be even clearer in their re-formulation a few years later as material rules of inference in the essay “Inference and Meaning” (IM). In IM, the synthetic purport of material rules of inference is clearly stated insofar as they are explicitly interpreted as conveying an extra-logical or material necessity (IM: §1), particularly of the kind which characterizes laws of nature. In this article we find Sellars contrasting his position with that of radical empiricists according to whom logical and inductive rules of inference are sufficient to account for synthetic knowledge (ibid.: §3), and causal inferences are ultimately reducible to “habitual expectations masquerading as inferences” (ibid.: §29). Firmly believing that the notion of material necessity had to be preserved, Sellars proceeds to show how logical rules alone are not, in the end, able to account for subjunctive conditionals, those special utterances that are so indispensable to the empirical sciences: he can thus conclude that “material rules of inference are essential to the language we speak” (ibid.: §26), so that the previous “Humean suggestion […] loses all plausibility” (ibid.: §29).

  • 16 “I am disturbed […] by the ease with which Lewis gets what he wants. One would have expected real c (...)

29If until now Sellars was talking from a strictly linguistic point of view, in “Concepts as Involving Laws and Inconceivable Without Them” (CIL) we find roughly the same topic examined from a modal point of view which explicitly deals with the problem of laws of nature. The highly technical (and hard to follow – Sellars was not exactly the smoothest writer, at least at the beginning of his career) 1948 paper is entirely devoted to a confrontation with Lewis’ solution to the problem of real connections, which they both understand as laws of nature. The starting point is to realize that the logical structure of a law of nature cannot be accounted for in terms of the traditional (x) (Fx –> Px): neither the classical material conditional, nor Lewis’ own conception of strict implication obtain. Although both agree on this, Sellars is not satisfied by Lewis’ alternative later solution of introducing a new logical connective specific to laws of nature.16 The rest of Sellars’ paper is thus devoted to an analysis of the distinctive kind of necessity characterizing laws of nature – an issue made unavoidable by Sellars’ “conviction that real connections of universals must be recognized in epistemology,” which he claimed to be the “most abiding of my philosophical prejudices” (CIL: §3), the only alternative being the dead-end of Humean skepticism. It is a problem that should by now sound familiar: namely, that of an extra-logical (natural, or physical) necessity which legislates for every actual and possible particulars of a given kind without being simply analytic.

30The framework settled on by Sellars is that of a possible world semantics constituted of possible histories, possible particulars and – this is where the novelty of his approach lies – possible universals. In order to restrict the infinite logically possible arrays of universals with particulars, thus enabling us to distinguish among different universals and different histories, the key element will be that of the common invariancies exhibited by the different histories. These invariancies are material insofar as they are “the only non-logical invariancies common to all these histories” (ibid.: §43) and are also “rooted in the universals [they] involve” (ibid.).

31Now, the highly technical nature of CIL is likely what prevents Sellars by providing examples, but it is possible to find one at the end of a similar train of thought presented in the slightly later “Language, Rules and Behavior” (LRB). Here, Sellars explicitly claims to be sketching “a regulist account of real connections and of the ‘synthetic a priori’ which preserves the insights of the rationalistic doctrine, while rejecting its absolutism as well as the pseudo-psychology of cognitive givenness on which this absolutism is based” (LRB: §26). He then proceeds:

It is important to note that the classical doctrine of synthetic a priori knowledge distinguishes carefully between the ontological and the cognitive aspect of such knowledge. Ontologically there is the real connection between the universals in question, say, Color and Extension. It is here that the necessity is located. On the other hand there is the cognitive fact of the intuitive awareness of this real connection, the Schau of the phenomenologist. Since it is a necessary consequence of the real connection of the universals that any exemplification of the one (Color) must also be an exemplification of the other (Extension), to see this real connection is to have rational certainty that the corresponding universal proposition “All colors are extended” will not be falsified by any future experience […]. (Ibid.: §27)

32In this passage, the relation between universals is exemplified through the concepts of color and extension: to state a real connection between the two means to state that there can’t be color without extension, that amounts to saying more than in point of fact all colors come with extension. The synthetic a priori proposition according to which “all colors are extended” (Husserl’s materially a priori proposition) implies that, although logically separable, colors and extension are materially inseparable: they are rooted together. The examples are precisely drawn off the Schlick-Husserl debate: “‘All colors are (necessarily) extended,’ ‘All tones have (necessarily) an intensity and a pitch,’ etc. The list is a familiar one” (ibid.: §39).

33By combining CIL’s modal analysis with the passage from LRB quoted above, it could easily look like Sellars has espoused Husserl’s notion of a material a priori almost unconditionally. Next to Carnapian formal rules of inference, he felt the urgence to introduce material rules of inference precisely to do justice to a kind of necessity not reducible to the logical one. Then, he supplemented his linguistic analysis by providing a modal, quasi-ontological analysis of the notion of real connection: that is why Quine, in reviewing CIL, asserted that here “modality is translated into ontology” (Quine 1949: 59). But a careful look is needed, as Sellars’ analysis does not stop here. The question asked by Sellars in LRB is indeed: how are we to account for the awareness of those real connections between universals whose existence is beyond doubt? Granted that “it is just as legitimate and, indeed, necessary for the philosopher to speak of real connections, as it is to speak of universals, propositions and possible worlds,” Sellars declares that, on the contrary, “it is just as illegitimate to speak of real connections as possible objects of awareness or intuition or Schau […] as it is to speak of apprehending universals, propositions and possible worlds” (LRB: §33). Sellars’ attack here is directed at the phenomenological doctrine of Wesensschau or “intuition of essences,” which he clearly does not consider a viable option.

34Although the account of the synthetic a priori provided here does not match Kant’s, Sellars is convinced that the “regulist position we are formulating could equally well be developed against a Kantian background, but that is a story for another occasion” (ibid.: 27n9). The last step I am going to present – concerning the pragmatic twist applied to the Kantian a priori – takes us seemingly even further away from the philosopher from Königsberg. For the moment I will not elaborate further on this statement, leaving the thoughts surrounding it to the end of the paper.

3. The Synthetic a Priori Revisited: The Pragmatic Twist

  • 17 As it is correctly pointed out in O’Shea 2018, the Kantian setting embraced by Lewis remains howeve (...)

35C.I. Lewis was not only a skilled logician but also a convinced Kantian. He is remembered as being the fiercest advocate of the teaching of Kant’s first Critique in North American universities (Brandom 2015: 2), and his best-known work Mind and the World Order (1929) is directly inspired by many Kantian assumptions. Beginning with the Preface, a strict connection between epistemology and ongoing scientific revolutions is stressed (Lewis 1929: vii): the exceptional advancements in logical analysis of the previous decades are those which, according to Lewis, enable us to clarify the analytic nature of all a priori truths – no longer restricted to the domains of logic and mathematics, but now extending to empirical concepts as well. Lewis, therefore, followed the then prevalent analytic trend in rejecting the notion of synthetic a priori knowledge, but at the same time – in a way similar to Schlick – enlarged the notion of analyticity to cover for a much wider range of cases, including what Kant conceived of as the Pure Understanding’s Synthetic Principles. All conceptual categories through which we interpret reality (Lewis calls them “criteria of reality” or “criteria of interpretation,” ibid.: 14) are on the one hand a priori, since their ground or warrant is independent of any particular a posteriori experience, but on the other hand they are analytic insofar as they are “implicit” in human experience and have been “brought to experience by the mind itself” (ibid.: 36). They correspond to the genuine import placed by the mind on what is given – a view that could be thought of as Kantian only if we set aside the fact that, according to Lewis, all principles of categorial interpretation are analytic rather than synthetic a priori.17 If the conceptual criteria of interpretation are always analytic a priori, empirical knowledge is, on the contrary, entirely a posteriori: “the assignment of any concept to the momentarily given […] is essentially predictive and only partially verified. […] Hence, all empirical knowledge is probable only” (ibid.: 37).

  • 18 A short version of Mind and the World Order’s main theses was reprinted in Feigl & Sellars 1949 und (...)

36I began this article by mentioning how issues related to conceptual change were the ones that most challenged Kant’s account since the 19th century. The idea of possible alternative conceptual frameworks coexisting with or superseding one another directly questioned the alleged universality and necessity of Kant’s synthetic a priori principles. Lewis was well aware of this and developed his theory of the pragmatic a priori accordingly:18 his conceptual pragmatism envisages precisely the existence of many different conceptual frameworks, all of which are useful for different purposes and ultimately chosen on pragmatic grounds. However, “[w]hether Lewis successfully accommodates both the unfalsifiability of any a priori conceptual scheme of laws, kinds, and realities in one sense (in virtue of their being analytically true by definition), but also in another sense the falsifiability of those empirical generalizations on which our knowledge of nature depends, is open to question” (O’Shea 2018: 209).

37In his 1953 paper “Is There a Synthetic a Priori?” (ITSA), Sellars explicitly tackles Lewis’ account of the pragmatic a priori and puts forward an alternative conception. His view on the matter had been anticipated at the end of several of the articles already mentioned, where Sellars pointed out a need regarding conceptual frameworks as wholes (CIL: §51; LRB: §43; IM: §48). The ending of IM is particularly appropriate because it explicitly refers to Lewis:

whereas he [the rationalist] speaks of the conceptual, the system of formal and material rules of inference, we recognize that there are an indefinite number of possible conceptual structures (languages) or systems of formal and material rules, each one of which can be regarded as a candidate for adoption by the animal which recognizes rules, and no one of which has an intuitable hallmark of royalty. They must compete in the market place of practice for employment by language users, and be content to be adopted haltingly and schematically. In short, we have come out with C. I. Lewis at a “pragmatic conception of the a priori.” (IM: §48)

38The need repeatedly pointed out by Sellars is the genuine pragmatist one according to which not only there are “an indefinite number” of alternative conceptual frameworks, but also that the choice for “one set of rules rather than another” (LRB: §43) – here “sets of rules” must be intended in the larger sense, as standing exactly for conceptual frameworks – is ultimately justifiable only on pragmatic grounds. And if this agrees with both – to the point that Sellars explicitly claim to have ended up with Lewis’ pragmatic conception of the a priori –, some frictions emerge right after:

Indeed, my only major complaint concerning his [Lewis’] brilliant analysis in Mind and the World Order is that he speaks of the a priori as analytic, and tends to limit it to propositions involving only the more generic elements of a conceptual structure (his “categories”). As far as I can gather, Lewis uses the term “analytic” as equivalent to “depending only on the meaning of the terms involved.” In this sense, of course, our a priori is also analytic. But this terminology is most unfortunate, since in a perfectly familiar sense of “synthetic,” some a priori propositions (including many that Lewis recognizes) are synthetic and hence not analytic. (Ibid.)

39This “terminological disagreement” between Lewis and Sellars is carefully analyzed by the latter in ITSA. Sellars begins by disambiguating some definitions: instead of the too vague sense according to which “analytic” stands for propositions “true by virtue of the meanings of the terms involved” (ITSA: 298-9) – Lewis’ sense –, Sellars decides to adopt Schlick and Husserl’s narrower definition for which the term “analytic” applies only to logically true sentences (e.g., truths of logic or their reformulations). On the other side, “a priori” is understood by him as a synonym for “true ex vi terminorum” (or “true by definition,” ibid.: 301), whereas Lewis defined it as “holding of all possible objects of experience” in a certain conceptual framework (ibid.: 300). The question that gives the title to the article can thus be reformulated into: are there propositions which are both synthetic and true ex vi terminorum?

  • 19 The reason is clearly explained by Sachs: “1) concept empiricism entails that every concept of a un (...)

40The path followed by Sellars to demonstrate that there are such propositions comes across in various steps, the first of which is the rejection of what he calls “concept empiricism,” that is, the assumption that conceptual content can be derived from sense experience: “only if concept empiricism is rejected is it possible to hold that there are non-logically true propositions which are true ex vi terminorum” (ibid.: 308). As a matter of fact, concept empiricism denies the existence of those real connections that synthetic a priori propositions establish in form of entailments between universals.19

41Concept empiricism also faces a major impasse concerning its inability to explain the acquisition of a conceptual framework in first place, to bypass which its proponents usually assume a theory of a direct awareness of abstract entities as a precondition of learning the intelligent use of symbols (ibid.: 312-3).

42Now, Sellars developed his functional-role semantics precisely in order to avoid this kind of pitfall, and here is where finally everything comes together:

The above dialectical examination of concept empiricism has been so designed as to bring me to the position I wish to defend, a position which […] represents a meeting of extremes, a synthesis of insights belonging to the two major traditions of Western philosophy, “Rationalism” and “Empiricism.” Stated summarily, it claims that […] the conceptual status of descriptive as well as logical – not to mention prescriptive – predicates, is constituted, completely constituted, by syntactical rules. […] The conceptual meaning of a descriptive term is constituted by what can be inferred from it in accordance with the logical and extra-logical rules of inference of the language (conceptual frame) to which it belongs. (Ibid.: 317)

43By embracing an inferentialist view according to which the meaning of a word lies in its norm-governed functional role in language – in other words, in rules governing its usage – Sellars made any ontological claim finally drop. Very little remains of the initial quasi-ontological analysis (“What, then, is the truth about real connections?” he asks emphatically towards the end of LRB. The answer: “Our use of the term ‘necessary’ in causal as well as in logical contexts is to be traced to linguistic rules,” LRB: §34).

44The now familiar logical and extra-logical (material) rules of inference are explicitly recalled by Sellars right after, and material rules of inference are precisely at the core of his re-interpretation of the problem of synthetic a priori:

Let me make the same claim in still another way by pointing out that where “x is B” can be validly inferred from “x is A,” the proposition “All A is B” is unconditionally assertable on the basis of the rules of language. Our thesis, then, implies that every primitive descriptive predicate occurs in one or more logically synthetic propositions which are unconditionally assertable – in short, true ex vi terminorum. But a logically synthetic proposition which is true ex vi terminorum is, by the conventions adopted at the opening of the chapter, a synthetic a priori proposition. (Ibid.)

45Here we are with Sellars’ complex, revised theory of synthetic a priori knowledge. How many synthetic a priori propositions are there? To Sellars, they are as many as material inferences and span from the inference from “x is blue” to “x is colored” (logically invalid, yet clearly a sound one) to causal laws themselves.

46Now, Sellars is aware that this theory of synthetic a priori propositions could not entirely satisfy its most strenuous defenders. “Is the synthetic a priori described above a real synthetic a priori? Would those who have fought and suffered for the cause of the synthetic a priori […] welcome me to their ranks?” (ITSA: 317). The matter is ultimately a “terminological decision,” depending on accepting the definitions Sellars gave at the beginning. To frustrate the radical synthetic a priori fringe is without doubt Sellars’ claim that “our conceptual frame is only one among many possible conceptual frames” (ibid.: 318). This would probably make them feel like “our synthetic a priori is a peculiar kind of a posteriori(ibid.). This, however, is also the big enhancement Sellars gives on Lewis’ theory of pragmatic a priori. Because not only we can, in a sense, carve out a space for synthetic a priori principles in forms of material principles of inference. And not only these principles must be conceived as ultimately answerable to experience and, eventually, subject to abandonment – something that also Lewis would have granted: the conceptual framework itself must also be conceived in this way. The paper can thus end, in the usual prosaic Sellarsian style, claiming that

[W]hile every conceptual frame involves propositions which, though synthetic, are true ex vi terminorum, every conceptual frame is also but one among many which compete for adoption in the market-place of experience. (Ibid.: 318)

4. Conclusion

47The path I have set out started with material rules of inference and ended with material rules of inference again. This likely didn’t come as a surprise to the habitual Sellars’ reader who knows about the profound systematic and holistic ambitions at the bottom of his philosophical enterprise. It is thus quite natural that concepts of causal necessity, synthetic a priori, real connections and material inferences are linked one with another, being, in a certain sense, different aspects of one system where tout se tient. We thus find out that those real connections traditionally expressed by synthetic a priori propositions are, in turn, regulated by conformation rules of languages. Sellars has, in the end, made real connections entirely immanent to thought (LRB: §40).

48Although initially it seemed that Sellars was proposing something very similar to Husserl’s theory of the material a priori, it slowly became clear that his solution to the problem eventually consisted in a linguistic, metaphysically uncommitted conception. Undoubtedly, he took the problem of the material a priori very seriously, but the solution for which he opts in the end shows how deep was the impact that Farber had on him: indeed, the version elaborated by Sellars in the end is in line with Farber’s project of a naturalization of phenomenology, to which the pragmatic twist on the synthetic a priori is added, so to speak, as the final piece. In this sense, Sellars’ distinctive defense of synthetic a priori knowledge is a place where it becomes evident how it is possible to philosophize across different traditions.

49Although all this seems at first glance very distant from Kant, there are at least two reasons why it is not. The first corresponds to the sense in which material rules of inference can be considered as a profoundly revisited and modified naturalist-adapted replacement for Kantian synthetic a priori principles. It has been rightly pointed out that what Sellars finally come up with “is a conception of cognition that in certain respects more closely resembles (than does Lewis’ view) Kant’s conception of synthetic a priori principles and of the ‘Copernican’ conformity to experience of our cognition, though again within a pragmatic and framework-relative conception of the a priori” (O’Shea 2018: 215). This is especially true if we consider that those extremely general principles (like the ones concerning “language games” of colors and sounds) are constitutive of our conceptual framework. The extra-logical rules of inference there at work cannot be a posteriori, insofar as they correspond to the rules which makes judgements possible. Yet, they are both a priori and revisable in light of experience.

50The second reason concerns a meta-philosophical point of view. In the course of the article, I briefly remarked on Sellars’ education showing how he was, in a certain sense, himself the hybrid product of distant philosophical traditions. In the early essays I have analyzed, the stereoscopic view he was looking for was represented as the via media between classic rationalistic themes and Humean empiricism. It is precisely in transitioning from his pure pragmatics’ essays to LRB, though, that the oversimplified clash between rationalism and empiricism begins to transform into a more sophisticated attempt to develop “a sound pragmatism” which “reject[s] descriptivism [here standing for radical empiricism] in all areas of philosophy” (LRB: §5). In “Language, Rules and Behavior” – where the formalistic approach is for the first time left aside – Sellars repeatedly contrasts the theory that I have labelled as a defense of synthetic a priori knowledge with the phenomenological Schau which pretends to intuit relations between universals. The position he will embrace at the end of the essay sides with that corrected pragmatism he will label as regulism, where phenomenological intuitions are replaced with patterns of rule-governed linguistic behavior. The key to this approach will be given to Sellars once again by Kant through what he will later develop as his theory of concepts as rules. But, to quote Sellars, that is a story for another occasion.

Top of page


Brandom Robert, (2015), From Empiricism to Expressivism: Brandom Reads Sellars, Cambridge, Harvard University Press.

Brandt Stefan & Anke Breunig (eds), (2019), Wilfrid Sellars and Twentieth-Century Philosophy, New York, Routledge.

Bordini Davide, (2011), “Analitico, sintetico e a priori: Questioni di forma. Il dibattito tra fenomenologia Husserliana e empirismo logico,” Rivista di storia della filosofia, 66 (2), 205-30.

Coffa J. Alberto, (1991), The Semantic Tradition from Kant to Carnap. To the Vienna Station, Cambridge, Cambridge University Press.

Corti Luca & Antonio Nunziante (eds), (2018), Sellars and the History of Modern Philosophy, New York, Routledge.

De Santis Daniele, (2015), “Once Again, Gigantomachy: Prolegomena to a Reinterpretation of the Schlick-Husserl Quarrel on the Synthetic A Priori,” Archivio di filosofia, 83 (3), 161-75.

Feigl Herbert & Wilfrid Sellars (eds), (1949), Readings in Philosophical Analyses, New York, Appleton-Century-Crofts.

Ferri Michela Beatrice (ed.), (2019), The Reception of Husserlian Phenomenology in North America, in collaboration with Carlo Ierna, Cham, Springer.

Gironi Fabio, (2015), “What Kant Has Ever Done for Us? Speculative Realism and Dynamic Kantianism,” in Sarah De Sanctis & Anna Longo (eds), Breaking the Spell: Contemporary Realism under Discussion, Mimesis International, 89-113.

Husserl Edmund, (2001), Logical Investigations. Volume 2, London and New York, Routledge.

Kant Immanuel, (1999), Critique of Pure Reason, transl. by Paul Guyer and Allen W. Wood, Cambridge-New York-Melbourne-Madrid, Cambridge University Press.

Lewis Clarence Irving, (1929), Mind and World Order. Outline of a Theory of Knowledge, London, Scribner’s Sons.

Livingston Paul, (2002), “Husserl and Schlick on the Logical Form of Experience,” Synthese, 132 (3), 239-72. Online:

Manca Danilo & Antonio Nunziante (eds), (2020), “Realismo, Pragmatismo, Naturalismo. Le trasformazioni della fenomenologia in Nord America,” Discipline Filosofiche, 30 (1).

Miraglia Roberto, (2006), “Dove iniziano gli a priori materiali? Schlick, Wittgenstein e le radici di un equivoco,” in Roberta Lanfredini (ed.), A priori materiale. Uno studio fenomenologico, Milano, Angelo Guerini e Associati, 95-120.

Nunziante Antonio, (2018), “Wilfrid Sellars e la matrice fenomenologica del dato: Farber, Husserl e il progetto di una naturalizzazione della fenomenologia,” Verifiche, 42 (1-2), 149-86.

Nunziante Antonio, (2020), “Marvin Farber e il progetto di una naturalizzazione della fenomenologia,” Discipline filosofiche, 30 (1), 135-57.

Olen Peter, (2016), Wilfrid Sellars and the Foundations of Normativity, London, Palgrave Macmillan.

O’Shea James, (2016a), “Concepts of Objects as Prescribing Laws. A Kantian and Pragmatist Line of Thought,” in Gabriele Gava & Robert Stern (eds), Pragmatism, Kant, and Transcendental Philosophy, New York, Routledge, 196-216.

O’Shea James (ed.), (2016b), Sellars and His Legacy, Oxford, Oxford University Press.

O’Shea James, (2018), “The Analytic Pragmatist Conception of the A Priori: C.I. Lewis and Wilfrid Sellars,” in Maria Baghramian & Sarin Marchetti (eds), Pragmatism and the European Traditions: Encounters with Analytic Philosophy and Phenomenology Before the Great Divide, London, Routledge, 203-27.

Pereplyotchik David & Deborah Barnbaum (eds), (2017), Sellars and Contemporary Philosophy, New York, Routledge.

Piana Giovanni, (1998 [1971]), “Husserl, Schlick e Wittgenstein sulle cosiddette ‘proposizioni sintetiche a priori’,” Aut Aut, 122, 3-30 [19-41]. Online:

Quine Willard Van Orman, (1949), “Sellars Wilfrid. Concept as Involving Laws and Inconceivable Without Them,” Journal of Symbolic Logic, 14 (1), 59-60.

Sachs Carl, (2014), Intentionality and the Myths of the Given: Between Pragmatism and Phenomenology, London, Pickering & Chatto.

Sanguettoli Filippo, (2020), “‘A Sympathy Laced with Disagreement.’ Wilfrid Sellars e la fenomenologia,” Discipline filosofiche, 30 (1), 235-59.

Scheler Max, (2014 [1913]), Der Formalismus in der Ethik und die Materiale Wertethik, Hamburg, Felix Meiner Verlag.

Schlick Moritz, (1949 [1931/1932]), “Is There a Factual a Priori?,” in Herbert Feigl & Wilfrid Sellars (eds), Readings in Philosophical Analyses, transl. Wilfrid Sellars, New York, Appleton-Century-Crofts, 277-85.

Sellars Wilfrid, (1947), “Pure Pragmatics and Epistemology,” Philosophy of Sciences, 14, 181-202. Reprinted in Sellars 2005. [PPE]

Sellars Wilfrid, (1947), “Epistemology and the New Way of Words,” The Journal of Philosophy, 44, 645-60. Reprinted in Sellars 2005. [ENWW]

Sellars Wilfrid, (1948), “Concepts as Involving Laws and Inconceivable without Them,” Philosophy of Science, 15 (4), 287-315. Reprinted in Sellars 2005. [CIL]

Sellars Wilfrid, (1948), “Realism and the New Way of Words,” Philosophy and Phenomenological Research, 8, 601-34. Reprinted in Sellars 2005. [RNWW]

Sellars Wilfrid, (1949), “Language, Rules and Behavior,” in Sidney Hook (ed.), John Dewey: Philosopher of Science and Freedom, New York, The Dial Press, 289-315. Reprinted in Sellars 2005. [LRB]

Sellars Wilfrid, (1953), “Inference and Meaning,” Mind, 62 (247), 313-38. Reprinted in Sellars 2005. [IM]

Sellars Wilfrid, (1963 [1953]), “Is There a Synthetic a Priori?,” in Wilfrid Sellars, Science, Perception and Reality, London and New York, Routledge, 298-320. [ITSA]

Sellars Wilfrid, (1968), Science and Metaphysics: Variations on Kantian Themes, The John Locke Lectures for 1965-66, London, Routledge and Kegan Paul. Reissued by Ridgeview Publishing Company in 1992. [SM]

Sellars Wilfrid, (1975), “Autobiographical Reflections,” in Castañeda Hector Neri (ed.), Action, Knowledge and Reality, New York, Bobbs-Merrill, 277-93. [AR]

Sellars Wilfrid, (1978), “Some Reflections on Perceptual Consciousness,” in Ronald Bruzina & Bruce Wilshire (eds), Crosscurrent in Phenomenology, The Hague, Nijhoff, 169-85. [SRPC]

Sellars Wilfrid, (2005 [1980]), Pure Pragmatics and Possible Worlds: The Early Essays of Wilfrid Sellars, edited and introduced by Jeffrey F. Sicha, Atascadero, Cal., Ridgeview Publishing Company. [PPPW]

Top of page


1 Citations to Sellars’ works will use the now standard abbreviations as listed in the References. For the articles collected in Pure Pragmatics and Possible Worlds citations will be to the numbered paragraphs that have also become standard in references to Sellars’ works, enabling cross-referencing across electronic editions.

2 Schlick 1949.

3 It goes without saying that not all logical positivists shared the same attitude towards the distinction: Carnap and Reichenbach were, in this sense, examples of an alternative approach. For a brief overview of theirs and other approaches, see Gironi 2015. For a more comprehensive outlook on Kant’s legacy within the analytic tradition, see Coffa 1991.

4 The majority of Sellars’ so-called “early writings” have been collected by Jeffrey Sicha in the volume Pure Pragmatics and Possible Worlds (2005 [1980]). The proposed time span for what count as “early” is merely conventional: another widespread option is to extend Sellars’ early phase to include all text published before EPM, thus stretching it until 1956.

5 The only close examination available for the early writings is Peter Olen’s Wilfrid Sellars and the Foundations of Normativity (2016). To say it with his words, “most of Sellars’ early publications are simply absent from the literature. Even though Sellars himself frequently cited some of his early publications, the papers that fall under ‘pure pragmatics’ make infrequent contributions to his overall philosophy” (Olen 2016: 4).

6 “The empiricism which I represent believes itself to be clear on the point that, as a matter of principle, all propositions are either synthetic a posteriori or tautologous; synthetic a priori propositions seem to it to be a logical impossibility” (Schlick 1949: 281).

7 To be sure, although Schlick’s article has been most often considered as a direct attack to Husserl, the main references to be found in the text are to Scheler’s Der Formalismus in der Ethik und die Materiale Wertethik. This led several scholars to claim that Schlick’s argument was affected by a substantial misunderstanding (Miraglia 2006; Livingston 2002) and that, with a more careful reading of Husserl’s texts, it would not have been impossible to rewrite Husserl’s argument in a form acceptable to Schlick (Livingston 2002: 248ff.). However, an in-depth analysis of the debate is far beyond the scopes of this paper, which is focused on Sellars’ take on the matter.

8 That they are not logical necessities can be appreciated by noticing that their terms can’t be substituted without impinging on their validity, insofar as their validity stems precisely from the material content we are talking about (color and sounds). In other words, their validity is determined by the kind of object they refer to: “the truths they enunciate are closely related to the particular species of content of experience (sound, color, spatial expansion) that constitute the reference of the enunciation” (Bordini 2011: 209; my transl.).

9 “Form (in the sense relevant to the debate) was, for both Schlick and Husserl, conceptual or logical; whereas to say of a proposition that it was ‘material’ meant that it depended on facts, intuitions, or the nature of experience. On the basis of this distinction, both philosophers agreed that propositions true in virtue of form are true a priori; their official difference, on Schlick’s construal at least, simply concerned whether there are further a priori propositions whose truth depends not on logical or conceptual form, but on the specific characteristics of experiential matter or worldly states of affairs” (Livingston 2002: 246).

10 For a more detailed comment on the matter see De Santis (2015: 172ff.).

11 The recognition of Sellars’ philosophy centrality and his admission among the “canonical” philosophers of the last century is proved by the flourishing of works in secondary literature dedicated to various aspects of his philosophy. Among the most recent are Brandt & Breunig 2021, Corti & Nunziante 2018, Pereplyotchik & Barnbaum 2017, and O’Shea 2016b.

12 For a deeper analysis of the relationship between Sellars and phenomenology, see Nunziante 2018 and Sanguettoli 2020. For a comprehensive overview of the history of phenomenology in the US, see Ferri 2019. For more on Marvin Farber, see Nunziante (2020: 151-7).

13 Even though this remains valid, it is also true that in the later production there are a few places where references to Husserl are massive and more explicit: I am talking about two of the talks collected in the Notre Dame Lectures, namely Perceiving and Perception (1973) and Scientific Reason and Perception (1977).

14 Part of Olen (2016)’s second chapter (esp. ibid.: 23-35) is devoted to explaining how the belief that Carnapian semantics was not able to account for extra-linguistic references stemmed from a misunderstanding inherited by Sellars from his colleagues (mainly Gustav Bergmann and Everett Hall) while working at Iowa University. In short, Bergmann and Hall interpreted Carnapian semantics as stipulating a relationship between two different language levels (the meta-linguistic level and the object-level), leaving the question of the relationship between language and world aside. However, Carnap’s shift to semantics was intended precisely to enable extra-linguistic reference, thus making the “Iowa interpretation” a substantial misunderstanding.

15 This quote must be approached with great care. The relations of “matters of facts” Sellars is here talking about remain, at this level, purely formal relations between predicates and individual constants. Although they are obviously fleshed out in our language, pure pragmatics “does not require formal languages to exhibit the factual relation between expressions and their extra-linguistic referents. Instead, pure pragmatics exhibit that such relationship must exist within any formal reconstruction of an empirically meaningful language, but not the actual relationships themselves” (Olen 2016: 44-5).

16 “I am disturbed […] by the ease with which Lewis gets what he wants. One would have expected real connections to be a bit more expensive, and the cry of ‘This or nothing!’ does not convince” (CIL: §3).

17 As it is correctly pointed out in O’Shea 2018, the Kantian setting embraced by Lewis remains however evident “in relation to the a priori conceptual grounds of our knowledge of nature. For insofar as any such knowledge […] is to be possible, for Lewis, some analytic categories of interpretation or other must always be presupposed as legislating a priori those laws that any real object of a given kind must obey in order to be a reality of that particular kind” (O’Shea 2018: 211). For more on how concepts of objects prescribe laws in Kant, Lewis, and Sellars, see O’Shea 2016a.

18 A short version of Mind and the World Order’s main theses was reprinted in Feigl & Sellars 1949 under the name “The Pragmatic Conception of the A Priori” (originally published in 1923). Lewis and Schlick’s articles were part of the section dedicated to synthetic a priori knowledge.

19 The reason is clearly explained by Sachs: “1) concept empiricism entails that every concept of a universal must be satisfied by some particular; 2) the concept of a real connection is the concept of a relation between universals, not between particulars; 3) so concept empiricism must deny that we can conceive of real connections” (Sachs 2014: 58).

Top of page


Electronic reference

Claudia Cavaliere, “Redefining the Synthetic a Priori”European Journal of Pragmatism and American Philosophy [Online], XIV-2 | 2022, Online since 17 October 2022, connection on 28 March 2023. URL:; DOI:

Top of page

About the author

Claudia Cavaliere

University of Padua

Top of page



Creative Commons - Attribution-NonCommercial-NoDerivatives 4.0 International - CC BY-NC-ND 4.0

Top of page
Search OpenEdition Search

You will be redirected to OpenEdition Search