Nelson Goodman’s Aesthetics—A Critique

Abstract

Nelson Goodman (1906–1998) is one of the leading American philosophers of the twentieth century. His well-known book Languages of Art is considered a major contribution to analytical aesthetics. While his views on particular issues have often been criticized, on the whole, he is considered to be a leading figure in twentieth-century aesthetics. Contrary to such a stance, I intend to argue that Goodman’s overall contribution to aesthetics is not as outstanding and valuable as is often maintained. Rather, I will try to show that his aesthetic views are grounded on a distorted representation of the earlier aesthetic tradition, without which they lose the novelty and originality ascribed to them. Once that representation is corrected, some of Goodman’s proposals turn out to be derivative and redundant. Additionally, where they do actually diverge from the earlier tradition and might stake a claim to originality, they turn out to be simply erroneous and misconceived, and sometimes even logically flawed. To conclude, Goodman’s lofty reputation as an aesthetician certainly requires major revision.

1. Introduction

Nelson Goodman is considered to be one of the greatest aestheticians of the twentieth century.1 Let us just quote just a few opinions:

Nelson Goodman has certainly been one of the most influential figures in contemporary aesthetics.

(Giovannelli 2017)

Nelson Goodman […] is also a famous figure in the analytic philosophy of art. […]

Goodman clearly ranks among the most important philosophers of the twentieth century.

(Cohnitz and Rossberg 2014, pp. 1, 2)

[…] in 1968 the third book emerged, Languages of Art, which was also to achieve epoch-making status.

(Ernst et al. 2009, p. 16)

It is hard to overemphasize the impact of Nelson Goodman’s (1968) classic, Languages of Art. The book offered a powerful new vision of aesthetics grounded in analytic philosophy of language, which reframed many of the questions being asked in aesthetics and gave original, ingenious, often eccentric answers to them.

(Robinson 2000, p. 213)

Admittedly, his views on particular issues have often been criticized. For example, in 1987, Douglas Arrell observed:

Nelson Goodman’s theory of pictorial representation is the best known and most widely rejected feature of his aesthetics. […] A survey of some forty of the articles and reviews which appeared in the wake of Languages of Art reveals that in about three-quarters of them this theory was a major topic of concern, and that overwhelmingly, the concern was to refute it […].

(Arrell 1987, p. 41)

Critiques of Goodman’s theory of pictorial representation include Kent Bach (1970), B. C. O’Neill (1971) and N. G. E. Harris (1973) and, after Arrell (1987), Peacocke (1987), Kulvicki (2003, 2006) and Blumson (2011).

Still, on the whole, Goodman is considered to be a leading figure in twentieth-century aesthetics, and criticisms of his views are sometimes seen as reflecting his originality and his questioning of received dogmas. At the same time, his views are also defended against criticism (see, e.g., van der Berg 2012), and discussed and analyzed as something cherished and valuable. For example, the editors of From Logic to Art: Themes from Nelson Goodman write in the Preface: “The contributions in this volume attest the fact that Goodman’s thinking holds many treasures waiting for being digged up.” (Ernst et al. 2009, p. v).

Contrary to such a stance, I intend to argue that Goodman’s overall contribution to aesthetics is not as outstanding and valuable as is often maintained. Rather, I will try to show that his aesthetic views are grounded on a distorted representation of the earlier aesthetic tradition, without which they lose the novelty and originality ascribed to them. Once that representation is corrected, some of Goodman’s proposals turn out to be derivative and redundant. Additionally, where they do actually diverge from the earlier tradition and might stake a claim to originality, they turn out to be simply erroneous and misconceived, and sometimes even logically flawed.

For my argumentation, in addition to Languages of Art (Goodman 1968), I will also refer to the article “Changing the Subject” (Elgin and Goodman 1987), which offers a concise presentation of Nelson Goodman’s aesthetic views (it was originally published in the leading American aesthetics periodical, Journal of Aesthetics and Art Criticism, in a Special Issue entitled Analytic Aesthetics).

2. “Traditional Aesthetics”

The first section of this article, headed “Incarceration”, represents an attempt at a reconstruction and criticism of the state of aesthetics before it became a focus of interest for analytical philosophy. At first glance, this reconstruction seems naively ahistorical, and the criticism appears to brand views that have long since lost their currency. The notion of “traditional aesthetics” in general, without any appellation, such as “traditional Enlightenment aesthetics” or “traditional aesthetics of German idealism”, suggests a vision of some supposed unchanging, homogeneous summa of views and ideas. The whole history of aesthetics is split into just two static stages. The first is “traditional aesthetics”, a term which turns out to mean little more than a collection of all the erroneous views which we would like to attribute to “others” and suitably criticize, and which are distinguished and gathered together partly so that our “correct” views might shine doubly bright in contrast with them. The second stage is analytical aesthetics, which, thanks to one single discovery, namely, the noting of a fundamental error underlying traditional aesthetics, resolves all paradoxes and misunderstandings. To put it another way, time is divided into a period of darkness before the coming of the prophet and the state following the revelation of the truth.2 Hence, Goodman evokes, for instance, Goya’s Disasters of War as evidence of the paltriness of “traditional aesthetics”, according to which all art is distinguished by beauty. However, he overlooks the fact that the category of beauty was deemed to be a defining characteristic of art in the mid-eighteenth century (it did occasionally appear in that role earlier), and by the time Goya painted this cycle (1810–1814), it had already undergone a process of transformation and supplementation (e.g., by the gradually emerging category of expression), precisely under the sway of changes in art and of such works as Disasters of War.

If, on the other hand, Goodman’s text were to be understood as questioning the usefulness of the category of beauty from a present-day perspective or from the point of view of the moment in time when analytical philosophy made its salutary and emancipatory entry into the fray, it must come across as a criticism of views which no one at that time was maintaining any longer. In the 1950s, when analytical philosophy took a more robust and unequivocal interest in aesthetics, beauty as a defining category of art was already only an historical phenomenon (Tatarkiewicz 1980, p. 23).

3. Anti-Essentialism—Yesterday, Today and Tomorrow

A more precise reading of Goodman weakens somewhat the initial impression of ahistoricity. A ripple in this static vision appears to be generated by his phrase “the pattern recurs”. This phrase sums up the emergence of the paradox of ugliness, “born of the conviction that beauty is essential to great art” (Elgin and Goodman 1987, p. 219), and introduces the invocation of another general category, that of “pleasure”, more psychological and historically later than “beauty”. So perhaps Goodman’s intention may be interpreted as follows: whenever we introduce any category with claims which go too far and overgeneralize, it will soon be undermined in the historical process by the reality of developments in art, and that which was perhaps initially intended as a generalizing description and the identification of crucial features will have to become merely a postulatory dogma. Understood in this way, Goodman’s argument would be levelled not against “beauty” or “pleasure”, invoked as historically grounded categories, but against the conviction “that aesthetic merit must derive from a single property common to all good art” or against the view that the aim of aesthetics is to seek “shared features that qualify objects, attitudes, experiences, and values as aesthetic” (Elgin and Goodman 1987, p. 219), which can only lead to failure and to the empty postulating of such properties.

Such a reading of Goodman’s position, which may be described as anti-essentialism and anti-fundamentalism invoking the argument of ineluctable historical change, essentially agrees in full with what is perhaps the strongest feature of analytical aesthetics. After all, during the 1950s, it attracted attention precisely with its sudden questioning of all generalizing attempts to capture the specific features of art, postulating that it was not just difficult to give any defining determinants of art, but on principle impossible, and that any attempts in that direction represented the fundamental error which underlay traditional aesthetics (see Weitz 1956; Kennick 1958).

Criticism and a dose of skepticism with regard to fundamentalism and to sometimes empty speculation is undoubtedly a positive phenomenon. Yet, such criticism—here from the perspective of analytical philosophy—is by no means something new, as Goodman appears to present it:

With the analytic turn, philosophy abandons the attempt to police shifting and inconsequent boundaries. It reconceives philosophy’s projects, recognizing that understanding neither begins nor ends with absolutes.

(Elgin and Goodman 1987, p. 220)

Yet, the supposition that it was analytical philosophy which for the first time in history questioned essentialist thinking and recognized “that understanding neither begins nor ends with absolutes” is entirely outlandish. Its mere formulation reveals its total fallacy. In fact, throughout the history of philosophy, essentialism has always been accompanied by a parallel anti-essentialism, to mention but its early incarnation in the philosophy of the sophists, opposed to Plato’s essentialism.3

4. The Duality of Emotion and Cognition in… Neopositivism

Another feature of “traditional aesthetics”, according to Goodman, is recognizing the strictly disjunctive duality of emotion, on one hand, and reason and cognition, on the other, and linking art and the aesthetic to the former domain; that is a diagnosis which must arouse at least surprise. To refer again to the “dawn” of philosophy: with Plato, the experiencing of beauty—and so aesthetic experience, as we may put it in considerably later terms—simply is at the same time the cognition of real being, that is, a cognitive experience. Additionally, according to the father of “traditional aesthetics”, Alexander Baumgarten (after all, Goodman, in using this term, clearly refers in part to the legacy of the eighteenth century), that discipline was the study of… cognition in its sensory variety. Additionally, although the development of aesthetics during the eighteenth century did not follow Baumgarten’s project exactly—since aesthetics soon began to mean that which it signifies for us, that is, the study of art and beauty—his cognitive orientation, also present in his views on art, became at least widespread during the eighteenth century. By way of example, we may invoke the perception—frequent during the eighteenth century—that pure instrumental music stood lower in the hierarchy than vocal music (and also, a fortiori, than other art forms), precisely on the grounds that it is lacking a signifying, rational, notional element. As Kant puts it, this is art which

[…] speaks by means of mere sensations without concepts, and so does not, like poetry, leave anything over for reflection […]. It is, however, rather enjoyment than culture […]; and in the judgement of Reason it has less worth than any other of the beautiful arts.

(Kant [1790] 1914, § 53, p. 217)

This judgement is grounded, of course, in the rationalistic eighteenth-century conviction that the cognitive and rational element was essential for something to be considered art of full worth, and its supposed lack prompts Kant to suspect that music lies outside the realm of culture, and so outside art. Moreover, with the advent of romanticism, that cognitive value would soon be ascribed to music as well, which began to be perceived as capable of reaching and reflecting the very (spiritual) essence of being, this time precisely because it is non-conceptual (see Schopenhauer [1819] 1909, vol. I, § 52; vol. III, chp. 39) Thus, Goodman’s suggestion that traditional aesthetics sharply contrasts emotion with cognition, and considers art to be wholly non-cognitive, simply contradicts the historical facts.

Moving forward, the perception of the mutual relations between emotion and reason, feelings and cognition, and their role in art assumed various configurations. Yet, one may hazard the assumption that a sharp opposition between these two domains and the ascribing of just one of them to art possibly had its locus classicus in one of the currents existing at the dawn of analytical philosophy, namely, neopositivism, which we certainly cannot assign to traditional aesthetics:

Many linguistic utterances are analogous to laughing in that they have only an expressive function, no representative function. Examples of this are cries like “Oh, Oh,” or, on a higher level, lyrical verses. The aim of a lyrical poem, in which occur the words “sunshine” and “clouds,” is not to inform us of certain meteorological facts, but to express certain feelings of the poet and to excite similar feelings in us.

(Carnap 1935, p. 28)

Carnap’s statement appears to suggest at the same time motivation for such a view on art: if cognition is to be a fully transparent domain of assertions, that is, statements satisfying the familiar neopositivist criterion of verifiability, free from metaphysical and any other muddiness, we must of course place art, with all its ambiguity and ineffability, entirely outside the sphere of the only true—that is, verifiable—cognition. Then, it is natural to link it exclusively to the non-cognitive domain of feelings and emotional expression. In other words, if we narrow down—in a neopositivist way—the domain of the cognitive, then art will find itself entirely outside that domain and turn out to be utterly non-cognitive. So is it that Goodman, protesting violently against the perception of art as wholly non-cognitive, is seeking above all to correct the mistake made by his predecessor (Goodman’s Ph.D. thesis, A Study of Qualities, was an analysis and elaboration of Rudolf Carnap’s work; see Cohnitz and Rossberg 2014, p. 4) but, out of consideration for his good name, ascribes it to some undefined “traditional aesthetics”?

5. Symbol, Density, Syntactic Differentiation

So what remedy does analytical philosophy—or essentially Goodman himself—propose for all these errors and woes of “traditional aesthetics”? Well, he suggests discussing art in terms of symbols, analyzing its objective meanings and references. The suggestion that such an approach to art is innovative belongs to the same set of rhetorical means as setting oneself up against some supposed “traditional aesthetics”. Given that even the greatest thinkers have resorted to this rather unsophisticated rhetorical device (to mention but the best-known example, perhaps, of Descartes, whose famous argument leading to cogito ergo sum can be found already in Augustine), we might well dismiss the whole thing with a shrug were it not for the circumstance that in this case, the lack of originality displayed by such an approach is all too stark. After all, interpreting art in terms of meanings and symbols is one of the most widespread of twentieth-century traditions, long predating the emergence of analytical aesthetics: suffice it to mention Ernst Cassirer (1923–1929), for example, as well as Hermann Kretzschmar (1902) and Arnold Schering (1914, 1941), two representatives of musical hermeneutics, and finally Goodman’s compatriot Susanne Langer (1942), who was roundly criticized by analytical philosophers. Additionally, this kind of approach to art, albeit without the direct use of the terms “sign” or “symbol”, is implicitly contained in many earlier aesthetic concepts, such as the Baroque theory of the affects, in accordance with which particular musical figures were objective bearers (signs, we would say today) of specific affects.

Of course, the use of a traditional notion in itself cannot be levelled as an accusation at philosophical theory. We should rather ask to what advantage that notion is applied. Some technical terms used by Goodman require a few words of explanation. “Syntactic density” and “syntactic differentiation” (Elgin and Goodman 1987, p. 221) are certainly among the core notions in Languages of Art. They are technical equivalents of the familiar opposites continuous–discrete and analog–digital. In the symbol system in which temperature is represented by a column of mercury, even the smallest change in the height of that column corresponds to a change in temperature—hence a column of any length may be understood as a different symbol. This kind of system of representation (a symbol system) is normally called analog or continuous, while Goodman calls it dense. In contrast, an alphabet, a set of digits or all natural numbers are discrete, discontinuous systems, in which each symbol is a “separate island”, incapable of passing smoothly to another. There are no intermediate symbols between individual letters, just as there is nothing intermediate between 1 and 2, 2 and 3, etc.

According to Goodman, vague intuitions resulting from such examples are not sufficient grounds for precise analysis of the different symbol systems that we can encounter in art, science and other fields of human activity. In chapter 4 of Languages of Art, entitled “Theory of Notation”, Goodman spares no pains to rigorously define these and associated notions with the use of a huge amount of formal technical means. Unfortunately, this turns out to be a spectacular flop: the terms introduced in a manner seemingly marked by meticulousness and formal rigor prove unfit for the purposes and applications which Goodman intends them for, and sometimes they are even defined in a logically incoherent way. Goodman’s explication of the analog–digital distinction was questioned by David Lewis (1971) and John Haugeland (1981), among others, and more recently by Corey J. Maley (2011), Aldo Frigerio, Alessandro Giordani and Luca Mari (Frigerio et al. 2013) and Matthew Katz (2016), while the notion of density itself is undermined by Guczalski (2022). Hence, his project for distinguishing between dense and differentiated symbol systems stands at the point of departure, that is, on the level of intuitions resulting from the examples given above. In effect, the technical terms that he uses may be safely understood on the basis of those intuitions, since Goodman’s formal definitions do not offer anything more precise—on the contrary, as essentially erroneous and wide of the mark, they may rather engender notional confusion.4

6. The Paradoxes of the Pan-Linguistic Vision

Space restrictions preclude a discussion here of the most important functions—as presented in Languages of Art—discharged by this notional distinction in Goodman’s aesthetic system.5 However, it is worth drawing attention to one of those functions, as it is linked to some highly controversial statements in the article under discussion. For Goodman, the density of a symbol system determines its representationality. A symbol is a representation insofar as it functions within a dense symbol system. It is not a representation if it is part of a discrete, or differentiated, system. From the relatively obvious observation that the similarity between two objects is not enough for one of them to represent the other (and consequently there must exist some other factor behind the existence of a symbolic relationship), Goodman draws the extreme conclusion that representation has nothing whatsoever in common with similarity or analogy.6 That inevitably leads to the conclusion that the meanings of such symbols are solely a question of arbitrary, conventional attributions. Essentially, this reasoning—erroneously leading from the observation that in representational symbols we are dealing with conventions to the conclusion that everything in them is convention—is the natural equivalent of the previous argumentation, erroneously leading from the observation that similarity is not everything for those symbols to the conclusion that it is nothing for them. Consequently, all the meanings of representational symbols become solely a question of conventionally established denotations—as in the case of language. Thus, Goodman’s whole model of symbol systems is based—in this sense—on the paradigm of language, contrary to his conviction that the criterion of density versus syntactic differentiation distinguishes non-linguistic from linguistic systems.

Indeed, it turns out that Goodman invariably thinks of all symbol systems in terms of the language paradigm. This is evidenced by such statements as the following:

We can no more tell what an unfamiliar work means “just by looking” than we can tell what an alien utterance means “just by listening”. To interpret a symbol correctly requires mastering its symbol system(s). And mastery is not given in the apprehension of the symbol.

(Elgin and Goodman 1987, p. 222)

That is perfectly true, one may say, but mastering a conventional language system is something completely different to “mastering” the symbol systems that we find, for example, in music or in painting, so even speaking of “mastering” them does not seem entirely appropriate. Taken literally, Goodman’s thesis says, for example, that we understand as much—or as little—by looking at a Japanese picture of a bird or a waterfall as we do by listening to a poem recited in Japanese.7 In other words, without mastering the symbol system—of Japanese art or language—we have no chance whatsoever of discerning whether this picture or any other depicts a bird or a waterfall, just as we have no chance of understanding that one word or another spoken in Japanese means “bird” or “waterfall”. Although such conclusions may seem absurd, that is precisely what results from Goodman’s views on representation. Indeed, in Languages of Art, we read, for example:

Realistic representation, in brief, depends not upon imitation or illusion or information but upon inculcation. Almost any picture may represent almost anything; that is, given picture and object there is usually a system of representation, a plan of correlation, under which the picture represents the object.

(Goodman 1968, p. 38)

To the objection that representation simply looks a little like the represented object, Goodman offers the following riposte:

Representational customs, which govern realism, also tend to generate resemblance. That a picture looks like nature often means only that it looks the way nature is usually painted.

(Goodman 1968, p. 39)

In short, representation does not rely on similarity. On the contrary, it is similarity that results from representation.

The paradigm of translating from one language to another appears also when Goodman writes about interpretation:

Even if their [critics’] readings are usually right, expertise does not make for rightness. Like skilled translators, astute critics may overlook an ambiguity, slight a subtlety, neglect a nuance, and so misinterpret a work.

(Elgin and Goodman 1987, p. 222)

Hence, the interpretation of works of art—poetry, music or painting—would be analogous to translating, for instance, from Japanese into English; it would be their translation into a verbal language. This way of perceiving art is sometimes treated—quite rightly—as one of the exemplary errors in thinking about art. For what does it imply? When we have at our disposal a correct, adequate translation from one language to another, for most needs and purposes it stands in for the original which in that sense becomes superfluous. This feature of the replaceability of some linguistic signs by others—which means that the function of perception of the actual signs themselves is purely instrumental—is linked to a property of language traditionally called semantic transparency: signs direct attention to their meanings, not holding attention on themselves. Based on the analogy between translating from one language into another and interpreting a picture or a musical work we would be obliged to say that a linguistically expressed interpretation essentially replaces a picture or a composition, that works of art are translatable into verbal language and that the function of their perception is purely instrumental, since they are replaceable by anything that can transmit the same meaning,8 such as an adequate interpretation. One would like to use Goodman’s words: “All this naturally follows from reasonable premises” (Elgin and Goodman 1987, p. 219)9 once we allow that all symbol systems are based on conventionally ascribed denotations.

Interestingly, in this case, Goodman realizes that such conclusions are inadmissible, as he notes elsewhere:

Interdisciplinary confrontations may be equally informative. We find that science scorns vague, ambiguous, and imprecise symbols; art welcomes them. In science, symbols normally refer singly and directly; in art, reference is often complex, multiple, and indirect. Scientific symbols are fairly attenuated; aesthetic ones, relatively replete. Science thus seeks nearly invisible windows through which its objects can be clearly discerned. Art tends to focus on symbols themselves.

(Elgin and Goodman 1987, p. 221; my italics)

He thus expresses, in slightly different language, what many others have said before him. Yet, the theoretical system is too dear to Goodman (as indeed it is to many other philosophers) to be modified if it contradicts the facts in an obvious way, as Goodman notes himself.

This imperialism of the theory of symbol systems manifests itself also in such statements as “Noting, for example, that both computer languages and musical notation are digital, we might test for correlations between the ability to write programs and the ability to write music” (Elgin and Goodman 1987, p. 221). One might just as well say that we can seek correlations between the ability to write poetry, cookbooks and books on physics, since they all use the same language.10 A lack of moderation and totalitarian temptations in applying reasonable theoretical distinctions and tools to elucidate matters with which they have little in common has already led to various ideologies—often more fraught with consequences than here.

7. Conclusions—Overcoming the Duality of Emotion and Cognition?

To close, a few words are due about how Goodman proposes to overcome that duality of emotion and cognition which he so fiercely brands and about the role he assigns to these two domains in relation to art. That overcoming is achieved by allocating a common task to cognition and emotion. Emotions in the perception of art are tolerated insofar as they can assist cognition, that is, insofar as they are instrumentally useful (a similar fate is shared by values):

Our revision of aesthetics displaces but does not disavow emotion. It takes the feelings works evoke, not as aesthetic ends in themselves, but as modes and means of understanding. Refined emotions, like discriminating perceptions, are aesthetically valuable because they enable us to discern and distinguish subtle but significant aspects of a work.

(Elgin and Goodman 1987, p. 221)

So merit, like emotion, transforms from end into means. We do not become connoisseurs to distinguish good art from bad; we learn to distinguish good art from bad to become connoisseurs—people who understand art, and through art their worlds.

(Elgin and Goodman 1987, p. 222)

So if emotions are the means, then what is the end? Cognition. Cognition of what? Of works of art. What for? In order to become connoisseurs who understand art. Why should we understand art? So that through it we can understand our world. Unfortunately, this last proposal—the only one that could give an answer to the question of art’s raison d’être—does not appear to result from Goodman’s earlier exposition. On the contrary, that exposition appears rather to suggest that understanding works of art (being a connoisseur) is the ultimate goal in his vision of aesthetics, whereas other elements that would bear hope for an elucidation of the importance of art, such as emotions or values, are instrumentalized as merely helpful in cognizing works of art.

Corresponding to such a vision in relation to science would be the thesis that its goal is by no means cognition of the world, but cognition of scientific theories. Moreover, such an approach is consistent with Goodman’s extreme nominalism, according to which it is we alone who create the world—through scientific theories, for example, or works of art. There is no other world beyond them. So only those theories or works of art can be the object of cognition. This does not mean, of course, that we do not cognize the “rest” of our world, the real world beyond theories and works of art. Quite simply, according to nominalism, the so-called real world, in its full extent, is only present in them and accessible to us through them.

There is no need for us to enter into a fundamental dispute with Goodman over the legitimacy of nominalism. We can note, however, that in suggesting that we understand our world through works of art, Goodman gives no indication of which aspects, areas or elements of our world we cognize in that way. It would seem, however, that emotions, the role of which is explicitly limited in the perception of art to the function of a means and instrument for cognizing works of art, are at the same time excluded as an object of cognition. Additionally, given the lack of any indication as to which elements of our world are cognized through works of art and given the emphasis Goodman places on “discern[ing] and distinguish[ing] subtle but significant aspects of a work”, we are left with the impression that the fundamental aim of our contact with art is … to cognize works of art. Things which would seem crucial to us, being the object of our profoundest interest, such as our emotional life, our spiritual world, are degraded to discharging a merely instrumental function.11

Since according to Goodman the sole purpose of art is cognition, that means going from one extreme—perceiving science and art as sharing nothing in common and forming complete opposites (an approach that Goodman rightly criticizes)—to another, one that practically equates art with science. Additionally, suggestions that art differs from science in the different character of the symbol systems it employs can only be dismissed by stating that fields of science differ from one another in the same way, which does not mean that some of them cease to be sciences. Indeed, the project of cognizing and understanding works of art, which Goodman presents as the essence of art and of our contact with it, is known, in more traditional terminology, as knowledge (or learning) about art as it is practiced in such fields as art history, musicology and literary theory.

“Art for art scholars” might be the slogan of Goodman’s aesthetic system. It hardly needs arguing that such a system is far from acceptable or convincing.