Navigazione – Mappa del sito

HomeNumeri73Improvisation and ontology of art

Improvisation and ontology of art

Alessandro Bertinetto
p. 10-29


I aim at explaining the sense in which the notion of improvisation is important for the ontology of art. In the first part, I criticize the widespread assumption of the repeatability of a musical work without transformation of its identity and defend Conversational Improvisational Emergentism (CIE) as the specific contribution of improvisation to musical ontology: in an improvisation, values and meanings of what has been played constrain what follows and are themselves retroactively (trans)formed by what follows; likewise, the performing interpretations of musical works and traditions reinvent their meanings and evaluation criteria, responding to past interpretations and opening up (as well as binding) possibilities for future interpretations. In the second part, I extend the scope of my investigation to art more generally. By critiquing the principle of «no evaluation without identification» and referring to Peter Lamarque’s and Joseph Margolis’ views about art ontology, I propose a transformative theory of artworks, which is based on the thesis of the improvisational “nature” of artistic practices (a thesis that I conceive of as a particular form of CIE). Its core point is that evaluative and performative interpretations of artworks (re)shape creatively, and retroactively, the meanings and the flexible identities of artworks. Accordingly, the artworks’ meanings and identities emerge (and are (trans)formed) through the cultural improvisational interactions in which artworks participate.

Torna su

Testo integrale

1. Introduction

1In this paper I will defend the view that artistic practices are improvisational in character (which does not mean that each artwork is ontologically an improvisation). I call this thesis: the Improvisational Nature of Artistic Practice (Inap) and it consists of two complementary claims.

    • 1 Cf. Bertram 2014: 113-121; Bertinetto 2017a.

    The first one goes as follows: Artworks are to be conceived of in terms of improvisational performances that (try to) make sense. When stating this, I am not defending the view that all artworks are produced on the spot. This would be an absurd claim. Instead, I mean that artworks come out of the interaction between elements (forms, materials, and media) and emerge ex improviso out of interactions with other artworks and the artistic practices (traditions, genres, styles…) they belong to. In this sense, they are an expression of all these elements and thus (trans)form them. The internal relations between the elements of the artwork as well as the interaction between artworks and artistic practices are of an improvisational nature1.

    • 2 Cf. Bertram 2010.

    The second claim follows from the first: Artistic practices and categories (genres, styles, traditions…) are in flux and the ways that they are formed and transformed provide an example of normativity without fixed norms2. Artistic practices and categories are not closed boxes in which to place the individual works of art. Instead, they are shaped by means of situated and responsive (interpretations of) artworks and they too emerge ex improviso out of the dynamic relationships between artworks, artists, critics and the various parties involved in different forms of reception, interpretation and aesthetic evaluation.

2The argumentative strategy that follows represents the extension of the theoretical perspective developed in my book, Eseguire l’inatteso. Ontologia della musica e improvvisazione (Performing the Unexpected. Ontology of Music and Improvisation: Bertinetto 2016), which explores the relationship between improvisation and musical ontology, to artistic practices.

2. (Re-)performing the unexpected

3The principal aim of the aforementioned book is to hold two apparently mutually exclusive theses together. I argue instead that they are theoretically interdependent. The two theses are the following:

  • The ontological specificity of musical improvisation.

  • The paradigmatic character of improvisation for musical ontology.

4In order to understand how both theses are connected in a plausible and theoretically enriching way, it is first of all necessary to explain and discuss them separately.

2.1. The ontological specificity of musical improvisation

5Let us start with the first point: the ontological specificity of musical improvisation.

  • 3 There is a vast literature on this topic: see for instance Davies 2001, Kivy 2002: 202-250, Dodd 20 (...)

6It can be said with certainty that the main philosophical problem of musical ontology hinges upon the relationship between musical works and performances3.Synthetically, the problem, and its commonly accepted solution, can be summarized in four steps.

  • A musical work ordinarily has multiple performances, and it is manifested, instantiated, and repeated in them.

  • By tacitly assuming the “no entity without identity” principle (Quine 1969: 23), the mainstream conception requires that the multiple repetitions of the musical work by/in its performances do not change the substance of the work. To put it bluntly: although repeatable, the musical work does not suffer a loss (or transformation) of identity and remains the same through different performances.

  • This thesis is seemingly descriptive in character. However, it involves a rigid normative constraint which can be formulated as follows: the performances of a musical work must portray it correctly in order to repeat it faithfully, so as not to change its identity. This prevents the work from becoming a different entity.

    • 4 Cf. Bertinetto 2012a: 108-113.

    The commonly accepted explanation of the problem of how a musical work can be repeated in different performances without undergoing ontological transformations is the dichotomy between (ideal or normative) Type and Tokens4. A musical work can be repeatedly manifested through different performances the same way a single type is a metaphysical structure that can be multiply instantiated (correctly and incorrectly) in different physical tokens. Following an established convention, I call this view structuralism.

  • 5 I use here Puy’s 2018 (335) paraphrase of Bertinetto 2012a, which, with a pragmatic self-contradict (...)
  • 6 It is ironic that by denying this difference between improvisation and the performance of a composi (...)

7Setting aside complications, my point is simple: the mainstream solution of music ontology “does not offer an adequate explanation of the ontological nature of musical improvisation”5. Improvisation escapes the duality between work and performance. Although, as I have explained at length elsewhere (Bertinetto 2012a: 106; Bertinetto 2016: 69-73) improvisation is notex nihilo and is always a matter of degree, there is no improvisation without the coincidence of invention and performance. Furthermore, in the performing arts, setting aside the complication of recordings (on this see Bertinetto 2016: 161-188), musical improvisation implies the coincidence of process and product as well as the contemporaneity of performance and perceptual experience. Hence, in the case of listening to an improvisation, it makes sense to say that listeners directly witness not only musicians’ interpretive skills, no matter how creative their performances may be, but also their creative activity in shaping artistic content to the moment in a specific situation. Conversely, the most relevant part of the content of a musical work, so the type/token theorist would say, is already there present as a structural type6.

  • 7 Definitional properties are properties that are parts of the definition of the item and constitutio (...)
  • 8 I have discussed them in Bertinetto 2012a, 2012b and 2016.

8According to the view, which considers the coincidence of invention and performance as the definitional property of improvisation which distinguishes between an item’sessential and constitutive properties7, an indicative (albeit approximate) list of the ontological essential and constitutive properties of improvisation may be articulated as follows8.

  • 9 Of course, these are generally properties of events and, obviously, also performing interpretations (...)

(a) Essential features of improvisation9

  • Situationality & Situativity: Improvisation occurs here and now; it is agency ontologically bound to its contextual environment (as constitutive element of the action).

  • Singularity: its space-temporal conditions are parts of its being: no two improvisations are identical. Copies, imitations, repetitions are by definition not improvisations.

  • Irreversibility: corrections are parts of the process/product.

  • Authenticity: agents are authors of their performances.

(b) Constitutive features of intentional artistic Improvisation

    • 10 Briefly, exaptation means making the situation or the environment appropriate through and for the p (...)

    Responsivity: the capacity of good reaction and adaptation to (as well as exaptive10impact on) environmental conditions.

  • Interactivity: horizontal and vertical interaction among different agents (norms, factual/empirical conditions, cultural constructs, individual and collective subjects).

  • “Normativity without norms” (Bertram 2010): improvisation develops its own norms accordingly to the specific situation of its occurrence, so that its norms are not merely presupposed, but are articulated and realized in/through the practice.

9The properties listed above clearly indicate that the traditional ontology of music, based on the concept of musical work, is inadequate to account for this artistic practice. Musical improvisation is not a work that can be interpreted more than once, nor is itself a performance of a work. However, in order to explain the ontological properties of improvisation, structuralists tend to force it into the traditional account of the relationship between work and performance provided by the type/token dichotomy, thereby creating conceptual acrobatics.

  1. A first acrobatic act (performed by Kivy 1983) consists in conceiving of improvisation as a performing interpretation that discovers the work-type as it manifests it, tokening it for the first time.

  2. A second acrobatic act (performed by Dodd 2013) consists in maintaining that a structure like that of a jazz standard is fixed and immutable, although it can be repeated through improvisational interpretations.

10However, these acrobatics do not work.

(a) Kivy’s acrobatics do not work for the following reasons:

  • First, it conflicts with the normative requirement of the mainstream structuralist solution, in that the work-type discovered by an improvisational performance is not normative. Consider the following: Performances of a musical work potentially provide incorrect instantiations of the musical work. But how could the first improvisational tokening possibly be incorrect? Furthermore, it can be reasonably assumed that, in this sense, any interpretative performance of a musical work discovers a non-normative type with a single token. So this solution would not be specific for improvisation’s ontology (Bertinetto 2012a: 110-111).

  • Secondly, Kivy’s idea conflicts with the essential qualities of improvisation. While the type discovered by improvisation is detachable and ontologically independent from the space-temporal specific and concrete conditions of its generative tokening, improvisation is ontologically fragile, in that it is counterfactually tied to its generative conditions (Bertinetto 2016: 169).

(b) The type/token model is not apt for explaining the ontological relationship between a jazz standard and the improvisational interpretation of that standards. While an ontological structure like a type is repeatable without transformation in/through its tokens, the standard jazz tune is not repeatedly instantiated in different performance without transformation. Only particular realizations of the standard concretely exist. Standards are, as it were, abstractions of concrete particulars (the performances) and are modified by those concrete particulars (cf. Feige 2014). In other words, particular performances (trans)form the standard that, per se, is only the abstract fiction we refer to in our cultural and linguistic practices (cf. Bertinetto 2016: 300-302).

2.2. The paradigmatic status of improvisation for music ontology

11My strategy for solving the problem of the ontology of improvisation is radically different. It does not consist in trying to explain how improvisation fits the mainstream structuralist solution, e.g. ordinary analytical ontological categories. Moving from the idea that we need a different ontology that is better suited to explaining the musical experience, my strategy rather aims at reorienting musical ontology by leaning on improvisation as paradigmatic artistic practice. As such, my argumentative strategy requires two steps and a conclusion.

  1. The first step concerns the clarification of the fundamental conceptual distinction between music as concretely performed and the “constructionality” of the musical work which, abstracted from its performances, is a cultural construct, a “social object” (see Ferraris 2015). I will argue as follows:
    Music has performative nature, so it is real (as performed or played back) “here and now.” Conversely, musical works are flexible constructs we refer to in our cultural practices. They are (perceived as) actually there only in/through performances. Musical improvisation is essentially performative, since its process is the artistic product. Hence, due to the performative nature of music’s reality, it is musical improvisation (which is essentially performative), and not the musical work (which is a cultural construct), that is paradigmatic for the ontology of music.

  2. The second step is the explanation of the transformative and procedural structure (or “logic”) of improvisation. In order to elaborate the ontology of music based on the improvisation model, it is necessary to understand the dynamic and transformative structure of improvisation as an open and recursive processual system in which each event is potentially unforeseen, unpredictable and emergent. This undetermined status remains through the events that follow, (trans)forms retroactively the sense of the past and its significance is in turn retroactively (trans)formed by what happens next (see Bertinetto 2016: 263-276).

  3. Finally, my point is that precisely this transformative dynamic “structure,” which typifies improvisation is exemplary for the relationship between musical works and performances (as well as for the relation among different musical works: cf. Feige 2014).

(a) Let us now move to the first step:

12The thesis of the constructionality and flexibility of the musical work involves the rejection of two ideological dogmas (that the mainstream structuralist musical ontology misconceived as ontological basic truths) that miss musical works’ cultural origin and ontological “nature” as human artefacts and cultural constructs.

  • The first dogma is the view of musical works as metaphysically fixed entities with invariant identities.

  • The second dogma is the repeatability of the musical work without transformation of identity.

13I propose a different model against these dogmas to fit our practices better than mainstream ontological structuralism (which serves the unrealistic Werktreue Ideal). My argument for defending the ontological constructioned status and flexibility of musical works goes as follows.

  1. Only as performed, musical works are empirically and concretely real as music that is perceived and experienced (empirical observation).

  2. Performances are concrete events; musical works are not concrete events.

  3. Hence, performances cannot be repetitions of a musical work (which is not a concrete event). For, the definition of repetition implies that something can be repeated only by something of the same nature (otherwise it is not a repetition).

    • 11 I thank an anonymous referee for pushing me on this point.

    The relation between performance and musical work is not a relation of repetition. It is a relation of realization. Performances give reality to the musical work, which without the performance exists only as a social-cultural object that, in the tradition of written music, is indicated by scores: real concrete objects that are not the musical work, but provide instructions for its realization. In a similar analogy, the score is related to the performance just as an architectural plan provides instructions for the realization of a building11. The performance exists; the musical work, otherwise only a cultural construct, concretely exists only in/through the performance.

  4. The realization of the work through the performance is (trans)formative. By realizing the musical work, each performance takes a stance toward the musical work as cultural construct and interprets the score (if there is one), mostly taking into account other past renditions. Performances appropriate the cultural construct through the score and adapt it to the performing situation, which is specific in material, technical, historical, social, cultural, aesthetics, and artistic terms. This adaptation performs – to different degrees – a transformation of the musical work as cultural construct (Creativity of performances).

  5. Hence, the idea that performances should match musical works fails as a criterion for evaluating performances. Matching is passed off as a criterion for preserving the assumed ontological identity of types (or classes) through “correct” performances, but in fact it is the translation of the aesthetic-evaluative ideal of Werktreue, operating within the framework of the classical Western tradition. It is a cultural prescription (cf. Ridley 2003).

  6. (ergo) Musical works are ontologically flexible, because they are continually re-configured (transformed) by the performances that bring them into reality.

  7. Abstracted from performances, musical works exist only as cultural constructs (fictions) we refer to (also in virtue of the scores as instructions for their realisation) in our changing cultural practices.

    • 12 Actually, a mental reading of a score may result in a determination of the musical work. As an anon (...)

    As a result, the identity of the musical work cannot be determined independently of its performances12. A musical work’s properties (i.e. its identity) are specified by performances and evaluations in musical practices (including testimonies and documents: scores, composers’ and interpreters’ annotations, critical literature, recordings, transcriptions…). Performances, as interpretations, are evaluative; evaluations are performative, in that they generate the identity of the musical work. Hence, to express this idea with a slogan, the operating principle is here “No identity without evaluation! ”. Also, the distinctions between good and legitimate (or correct) performances and between essential and accidental elements of the musical work that seem to be a necessary condition for the identity of the musical work, depend on evaluations that may be performed and understood only in (reference to) particular musical performances and practices.

14To sum up the first step of my argument, I defend the ontological (trans)formability (flexibility) of musical works. This is due to musical works’ ontological identity depending on, and co-varying with, musical and aesthetic practices (and contexts). Since practices change, the identity of musical works is dynamic (in flux).

  • 13 We may accommodate this idea within the type/token ontology. In order to do so we should however ab (...)

15As a matter of fact, the standards of correctness and of (artistic) excellence for performances depend on musical practices. They are not established by an alleged metaphysical identity of musical works. Hence, if performing conventions and reception methods change and if they condition the identity of the musical work, then the identity of a musical work as a cultural construct can change: (different and changing) critical and performing interpretations affect artistic properties of the musical works (their identity)13.

(b) Here is the second step of my thesis.

16The key question at issue is the following. Why is improvisation paradigmatic for musically “flexible” ontology? The answer is to be found in what I call the transformative “logic” of improvisation.

    • 14 An open system develops its own limits (See Luhmann 1995) in an autopoietic way (see Maturana and V (...)

    Improvisation is a (trans)formative making in which process and product constitute a self-referential and reflective relationship (an open system)14.

  1. In open systems outputs are reintegrated as inputs and the process lives on by virtue of feedback (retroactivity). As a consequence, sounds (and silences) that are performed are not only outcomes of the production process, but feedback into the production process, further impacting the way the process will continue.

  2. Hence, as an open system, in improvisation a dialectic between plan and action is at work: improvisation – as opposed to the execution of a pre-established plan – takes place in a single phase: the plan is formed while being performed: this means, that also the (cultural, technical, normative, aesthetic, etc.) preconditions of improvisation (its background) are transformed in their effective development, what is produced becomes a new condition for the production process.

17Hence, the normativity of improvisation is shaped in/as performance (normativity is, in a way, improvised: cf. Bertinetto and Bertram manuscript). This can be explained by the following two claims:

  1. Improvisation generates normativity: Improvisation is not only acting on the basis of norms. In its unfolding, improvisation establishes norms (i.e. makes sense), in that each move constraints following moves.

  2. Normativity is improvised: The specificity of the situation remains unforeseen, but normativity should account for it. Thus, the application of norms (and evaluation standards) in a specific situation potentially involves, in a retroactive way, their (trans)formation. This relates to Wittgenstein’s famous sentence: “We make up the rules as we go along” (Wittgenstein 1953: § 83).

18Therefore, the unexpected and unprecedented are not only what is outside the norm, but also what the norm requires for its own application (which suspends and (trans)form the norm in order to copying with reality).

19Furthermore, even judgment about the value of improvisation is not guaranteed by the reference to pre-established criteria. The criteria for evaluating an improvisation are also put into play in the improvisation (both for public and for performers). In order to understand and evaluate an improvisation we cannot limit ourselves to using abstract criteria. The evaluation criteria are (re)generated through the performance, which suspends the normativity on which it is based, while it applies hic et nunc, thereby producing the unexpected.

(c) The conclusion of the argument follows: Due to both the transformative and retroactive normative “logic” of performance, improvisation is the model of musical practices. This claim can be explained as follows:

  • The normative force of a musical work (or of a musical genre) depends on the actual performing practices (cf. Davies 2012). In applying the norm (consisting in a musical work, in a musical practice, in a musical tradition), performances realize them in specific ways, adapting them to concrete situations; and this performative application of the norm potentially (trans)forms the norm.

  • As a consequence, the identities of musical works, musical practices, and traditions are ongoing construction(s): they are concretely real only in the performances that retroactively transform them.

20The thesis I call Conversational Improvisational Emergentism (CIE) results from the arguments made thus far. It goes as follows:

21The story of (performing and critical) interpretations of a musical work is the musical work as cultural construct. This social construct is an outcome which can be continuously transformed and acts as a kind of interactive conversation. Its meaning is not immutable as the result of a compositional act (eg. by author’s intentions): rather, it emerges out of cultural interactions.

  • 15 For this “Gadamerian” view of the relationship between musical works’ performances see Feige 2014.

22Each (performing and critical) interpretation produces meanings, thereby retroactively (trans)forming musical work’s meanings and identity; it is a contribution (non-programmable in advance) to further re-configuration of the musical work’s “fictional” identity. The “dialogue” among the interpretations of a musical work transforms the musical work and is open because it builds out of itself its own normative context15.

23To summarize: in an improvisation, values and meanings of what has been played constrain what follows and are themselves retroactively (trans)formed by what follows. Likewise, the performing interpretations of musical works and traditions reinvent their meanings and evaluation criteria, responding to past interpretations and opening up (as well as binding) possibilities for future interpretations.

3. Improvisation and the ontology of art

24The reflections carried out so far concern music ontology. The question that now arises is: Is it possible to generalize this approach?

25The answer is yes. Of course, my re-orienting strategy is applicable to other performing arts as well, where there is a relation between an allegedly repeatable artwork (for instance a theatrical work) and its realizations/manifestations (cf. Hamilton 2007).

26Furthermore, I venture that we may conceive of improvisation as paradigmatic for ontology of art as such. My general point is that a look towards improvisation helps us to overcome some of the objectivistic trends in contemporary ontology of art (what I mean by “objectivistic” will be clear in the next section) and conceiving of artworks as dynamic, i.e. as (trans)formable as well as (trans)formative entities. In fact, like elements of improvisations, artworks are continuously reshaped by the interaction to which they participate, thereby contributing, as kinds of agents, to (trans)form artistic practices and categories and, consequently, also other artworks’ identities.

3.1 No identification without evaluation

27As mentioned in the introduction, I call this thesis concerning the conceptual contribution offered by improvisation for the ontology of art, Inap: Improvisational «Nature» of Artistic Practices.

28Inap contrasts with largely accepted (but wrong) principle of (objectifying) art ontology, according to which the ontological identification of an artwork is a necessary pre-condition for its evaluation.

29Against this principle, Inap highlights the performative force of evaluation and the evaluative impact of performance, thereby accepting the principle: “No identification without evaluation”. This means that the ontological identification of an artwork is conditioned by evaluative/interpretive practices.

30The task of the rest of the article is to argue in favour of this thesis. I will defend it by extending CIE to art in general.

3.2. Lamarque’s Contribution

31I will begin by discussing the proposal made by Peter Lamarque. According to Lamarque (2010: 51), the artworks’ “identity conditions, being value-laden, are distinct from those of functionally-defined artefacts and physical objects in the natural world”.

32This depends on the following reasons:

  1. Artworks are products of human creativity (Lamarque 2010: 42) which “essentially involves bringing something new into the world” (Lamarque 2010: 46). This assertion contradicts the Platonist view of artworks, which holds that they are timeless universals and, therefore, are discovered and found but not produced or “created” by artists (cf. Kivy 2002: 202-223).

  2. Artworks need “a complex cultural background of practices” (Lamarque 2010: 41), “appropriate beliefs, attitudes, modes of appreciation, and expectations” (Lamarque 2010: 54) in order to (continue to) be the artworks they are.

  3. Practices, cultural contexts and their normative force are flexible and can change (even when they are highly structured: see Guala 2016).

33For these reasons, artworks can change (and even disappear) depending on cultural practices. As Lamarque (2010: 68) writes:

Nothing can be a work (of art) if it does not play a role, or be fit to play a role, in a human practice where a sufficient number of informed practitioners recognize its status and respond appropriately. […] Works cannot survive as works if these practices are lost. If no-one is any longer in a position to judge that something is a work of a particular kind then works of that kind no longer exist.

34Lamarque is right about that: artworks’ identity depends on the cultural practices to which they participate. However, Lamarque’s view is crucially flawed in that, following the view elaborated by Kendall Walton in an extremely important, but partly mistaken, article on philosophy of art (Walton 1970), he believes that artworks can be appreciated and evaluated properly only when they are placed in the right categories or genres (Lamarque 2010: 75). Briefly, the problem is the objectification of the relation between an artwork and its category: according to Lamarque, evaluations (i.e. value judgements) make ontological identification possible, but they do not make sense or generate categories (genres, norms and values).

  • 16 Indeed, as we have seen, normativity is improvisational, since the rules, to be valid, must be plas (...)

35One can briefly explain the flaw of Lamarque’s view by saying that, following Walton, he underestimates the performative potential of interpretive evaluations. In this way, he misses the dynamism of the normativity of human practices16 as well as the specific transformative nature of artistic practices, which constitutes one of the greatest contributions of art to human practices.

3.3. Margolis’ Answer

36To solve this problem, I will now turn to Joseph Margolis. His “hermeneutical pragmatism” (as I characterize his view) offers a valuable contribution to the question of art ontology.

37Margolis (1999) properly highlights the performative role of evaluative interpretations for art ontology. His argument can be explained as follows:

  1. First, like Lamarque, he defends, the grounding role of evaluation for individuating artworks. To individuate an artwork (as different from a physical object) is to assign it cultural/interpretive/intentional/evaluative “parts” and doing so belongs to the same practice of imputing meaning to the object by means of interpretation.

  2. Secondly, as opposed to Lamarque and Walton, Margolis explains that “you cannot settle the ontology of art by imposing a priori constraints on the logic of interpretation” (Margolis 1999: 95). Artistic categories are not preconditions for ontologically identifying artworks. They are also the historical results of interpretative and evaluative activities with which appreciators generate expectations regarding the production and the assessment of new artworks, thereby assigning ontological identity to artworks (cf. Condello-Terrone 2017: 41).

38Accordingly, artworks are (1) (trans)formable and (2) (trans)formative.

(1) Artworks are transformable, because, like persons, they are, in Margolis’ felicitous definition, “physically embodied and culturally emergent entities” (cf. Margolis 1974).

39Indeed, cultural entities are not objectively fixed or invariant. They differ from purely physical entities, in that the numerical identification of artworks does not necessarily require the description of their nature by means of assigning them determinate and invariant properties – i.e. attributing them predicates following bivalent logic (as pretended by neo-scholastic positions: cf. Pouivet 2017).

40The artwork’s identity can be specified only in terms of cultural properties that are in flux, in that they are subject to interpretive and evaluative transformations situated in a given cultural context and responsive to that particular context as well as to other interpretations. Interpretive transformations of this kind are rightly termed improvisations (cf. Margolis 1999: 96). They (trans)form artworks’ identity and are not constrained by fixed criteria of practices, genres and styles.

(2) Artworks are transformative, in that they are to be understood as performances impacting on evaluative criteria (cf. Noë 2015).

  • 17 Cf. Condello and Terrone 2017 (43): “As a historical outcome, a genre has an origin and it can unde (...)

41Artistic categories (genres, styles…) are not simply invariant objective properties that normatively qualify artworks. This is due to the fact that artworks actualize (realize) and (trans)form artistic categories: an artwork is not content that enters the category like an object we put in a box. In brief, artworks are transformative in that they intervene in reality, and include the determination of artistic categories, practices, genres17. Now, to understand the sense in which the work of art is transformable and at the same time transformative, one can resort to the concept of improvisation, arguing that the relationship between the work of art and the artistic category is improvisational. It is improvisational because the relation between artworks and artistic categories (genres, styles, traditions, etc.) is not a one-way relation of determination, but a mutual retroactive interaction. We can clarify this relationship by articulating its two main aspects.

  1. Artistic categories develop through (interpretative evaluations of) artworks, are (trans)formed through (interpretative evaluations of) artworks, emerge out of (interpretative evaluations of) artworks.

  2. Artworks are ex improviso (cf. Bertinetto 2017). While applying the category/norm, they help to (trans)form it. Thus, they cannot be reduced to pre-existing categories.

  • 18 Cf. Condello and Terrone 2017. In my view (which is openly evaluative and not classificatory) this (...)

42At this regard, Margolis’ example is telling. Referring to Picasso, he writes that “Picasso’s innovation” (with Les demoiselles d’Avignon) “cannot be routinely reconciled with any of the would-be canons of well-formed painting up to the intrusion of Les demoiselles […]” (Margolis 1999: 93). In other words, Picasso’s innovation – like every authentic or exemplary18 artworks – is a kind of improvisation on categories-norms and on other artworks, which is “open to further improvisation” (Margolis 1999: 94), thereby (trans)forming categories and norms.

  • 19 For the notion of signifying in relation to improvisation see Monson 1996; cf. Bertinetto 2017b.
  • 20 Thus, even granted Laetz’s interpretation of Walton’s paper (see Laetz 2010), according to which th (...)

43Improvisers base their performances on a background of artistic traditions and aesthetic/artistic materials, forms, techniques, styles and conventions, from which they improvise, re-signifying this background and thus bringing it to life19. Artists, in this sense, are all improvisers: they bring out something new from old materials and forms, using and abusing them for unexpected inventions and, in this way, bringing them a new life (cf. Bertinetto 2018). For instance, Picasso’s improvisation withLes demoiselles d’Avignon, signifies and performatively interprets and evaluates, (on) the Cezanne painting(s) Les grand baigneuses and on African ritual masks: Old artworks and artistic categories get new meanings and new identities through new artworks20.

4. Conclusion: Conversational improvisational emergentism21 (CIE)

  • 21 Notions like emergence and emergentism can be applied to the artistic field in different ways. For (...)

44We can now draw conclusions based on the reflections made so far. Artworks are cultural constructs generated by interpretive and evaluative meaning attributions made by different cultural entities (including persons): artworks have emergent identities.

45The ontological identity of an artwork results out of the dynamics of artistic practices: production, interpretation and criticism (cf. Bertram 2014). This dynamic is improvisational in character: its model is the interactional relation between musical works and performances as well as between different performances of a musical work.

  • 22 Cf. Bertinetto 2017a.

46Indeed, improvisational (interpretive, evaluative, and performative) interactions between different agents at different levels feedback into the artworks’ identities22. Hence, although the physical object may remain (or be restored as) the same, the emergent intentional (cultural) «part» is in flux and, consequently, the artworks’ identity is continuously (trans)formed through the way they are interpretatively and evaluatively interacted with and appropriated.

47Moreover, artworks are not simply passive subjects in these interactions. Rather, they actively participate in those interactions, so that they impact on artistic criteria of evaluation and interpretation, thereby contributing to (trans)forming themas well as other artworks. So it may be said that each artwork is a retroacting response to a specific situation, to artistic categories and to past artworks whose identities is transformed due to the performative evaluative interpretation enacted and embodied by the new artwork.

48This can be finally explained by the construction of a structural analogy with the dynamics of conversations. According to Sawyer 2001, the meaning of conversations is emergent. This amounts to saying that conversations are creative, in that their meaning cannot be traced back to the semantic meaning of the words used, but emerges pragmatically, and creatively, out of the conversational interaction (see Bertinetto 2017b: 14-15).

  • 23 In Bertinetto 2017b I resorted to this idea for explaining musical meaning.

49Analogously, the interpretation of artworks is a kind of conversational dialogue with artworks (Cf. Carroll 1992). However, like in real conversations, this dialogue does not aim at discovering the author’s intentions (Huddleston 2012). Conversations follow conventions, norms and constraints, but the specific conversational situation reshapes norms, as speakers’ reciprocal interactions impact pragmatically on the context (and on established meanings resulting from past evaluative interpretations).23 Hence, artworks’ ontological identity is conversational, since it is not reducible to what artists claim subjectively, but it is (trans)formed through the evaluative and performing interpretations it obtains in its cultural life.

50In conclusion, CIE explains INAP. As paradigmatically exemplified by the “improvisational”-(trans)formational link between musical works and their performances, evaluative and performative interpretations of artworks (re)shape creatively, and retroactively, the meanings and the flexible identities of artworks. Accordingly, the artworks’ meanings and identities emerge (and are (trans)formed) through the cultural improvisational interactions in which artworks participate. They cannot be simply explained by tracing them back to pre-established artistic normative categories, because they impact on those categories, (trans)forming and renewing them.

  • 24 Cf. Hallam-Ingold 2007.
  • 25 Previous versions of this paper have been presented in Paris on March 22nd 2018 (at the workshop L’ (...)

51To sum up, like in the exemplary musical case, nothing is ontologically stable in the artworld as both artworks’ identities and artistic categories emerge out of cultural interactions and are, therefore, continuously in flux24. This state of flux, in which each new outcome feeds back the process, retroacting on the identity of its elements, is typical of improvisation. As a consequence, improvisation is not an ontological rare beast: rather, it offers an explanatory model for the ontology of art25.

Torna su


Bertinetto, A. 2012a, Paganini does not repeat. Improvisation and the type/token ontology, “Teorema”, 31, 3: 105-126.

Bertinetto, A. 2012b, Performing the unexpected. Improvisation and artistic creativity, “Daimon”, 57: 61-79.

Bertinetto, A. 2016, Eseguire l’inatteso. Ontologia della musica e improvvisazione, Roma, il Glifo.

Bertinetto, A. 2017a, Ex Improviso, Trans-Formation als Modell künstlerischer Praxis, in K. Maar, F. Ruda, J. Völker (eds), Generische Formen. Dynamische Konstellationen zwischen den Künsten, München, Wilhelm Fink: 143-158.

Bertinetto, A. 2017b, Sound pragmatics. An emergentist account of musical meaning, “Rivista italiana di filosofia del linguaggio”, 11, 2 “The Languages of Art”: 1-29.

Bertinetto, A. 2018, The birth of art from the spirit of improvisation, “Quadranti”, 6, 1: 119‑147.

Bertinetto, A. 2019, L’emergentismo nell’arte, “Philosophy Kitchen”, 11, 7: 177-191.

Bertinetto, A., Bertram, G. (Manuscript), “We make up the rules as we go along” - Improvisation as an Essential Aspect of Human Practices?

Bertram, G. 2010, Improvisation und Normativität, in G. Brandstetter et al. (eds), Improvisieren. Paradoxien des Unvorhersehbaren, Bielefeld, Transcript: 21-39.

Bertram, G. 2014, Kunst als Menschliche Praxis, Berlin, Suhrkamp.

Carroll, N. 1992, Art, intentions and conversations, in G. Iseminger (ed.), Intention and interpretation, Temple University Press, Philadephia: 97-131.

Condello, A.,Terrone, E. 2017, Genre classification: a problem of normativity and exemplarity, in W. Gephart, J. Leko (eds), Law and the Arts. Elective Affinities and Relationships of Tension, serie del Käte Hamburger Center for Advanced Study in the Humanities Law as Culture, Werner Gephart (ed.), vol. 18, Frankfurt a.M., Klostermann: 39-46.

Davies, D. 2012, Enigmatic variations, “The Monist”, 95, 2, 1: 643-662.

Davies, S. 2001, Musical Works and Performances, Oxford, Clarendon Press.

Dodd, J. 2007, Works of Music. An Essay in Ontology, Oxford, Clarendon Press.

Dodd, J. 2013, Upholding standards: A realist ontology of standard form jazz, “Journal of Aesthetics and Art Criticism”, 72, 2: 277-290.

Feige, D.M. 2014, Philosophie des Jazz, Berlin, Suhrkamp.

Ferraris, M. 2015, Collective intentionality or documentality?, “Philosophy and Social Criticism”, 41, 4-5: 423-433.

Fischer-Lichte, E. 2004, Ästhetik des Performativen, Frankfurt a.M., Suhrkamp.

Giombini, L. 2017, Musical Ontology. A Guide for the Perplexed, Milano-Udine, Mimesis International.

Guala, F. 2016, Understanding Institutions, Princepton and Oxford, Princeton University Press.

E. Hallam, E., Ingold, T. (eds), 2007, Creativity and Cultural Improvisation, Oxford - New York, Berg.

Hamilton, J. 2007, The Art of Theatre, Oxford, Blackwell.

Huddleston, A. 2012, The conversational argument for actual intentionalism, “British Journal of Aesthetics”, 52: 241-256.

Kivy, K. 1983 Platonism in music: A kind of defense, “Grazer Philosophische Studien”, 19: 109-129.

Kivy, K. 2002, Introduction to a Philosophy of Music, Oxford - New York, Oxford University Press.

Laetz, B. 2010, Kendall Walton’s ‘Categories of Art’: A critical commentary, “The British Journal of Aesthetics”, 50, 3: 287-306.

Lamarque, P. 2010, Work and Object, New York, Oxford University Press.

Levinson, J. 1990, Music, Art, and Metaphysics, Ithaca, Cornell University Press.

Levinson, J. 2011: Indication, abstraction, and individuation, “Tropos. Rivista di ermeneutica e critica filosofica”, 4, 2: 121-131.

Luhmann, N. 995, Social Systems. Stanford, Stanford University Press.

Margolis, J. 1974, Works of Art as physically embodied and culturally emergent entities, “The British Journal of Aesthetics”, 14,3: 187-196.

Margolis, J. 1999, What, after all, is a work of art? University Park, Pennylvania State UP.

Maturana, H.R., F. Varela 1980, Autopoiesis and Cognition, Dordrecht, Reidel.

Monson, I., 1996, Saying Something. Jazz improvisation and Interaction, Chicago, The University of Chicago Press.

Noë, A. 2015, Strange Tools, New York, Farrar, Straus and Giroux.

Quine, W. V. O. 1969, Ontological Relativity and Other Essays, New York, Columbia University Press.

Pouivet, R. 2017, L’art et le desir de Dieu, Rennes, Presses Unviersitaires de Rennes.

Puy García Carrill, N. 2018, The Ontology of Musical Versions and Transcription, Doctoral Thesis, Universidad de Granada.

Ridley, A. 2003, Against musical ontology, “Journal of Philosophy”, 100: 203-220.

Sawyer, R.K. 2001, Creating conversations: Improvisation in everyday discourse, Cresskill (NJ), Hampton Press.

Walton, K. 1970, Categories of art, “The Philosophical Review”, 79: 334-367.

Wittgenstein, L. 1953, Philosophical Investigations, Oxford, Blackwell

Torna su


1 Cf. Bertram 2014: 113-121; Bertinetto 2017a.

2 Cf. Bertram 2010.

3 There is a vast literature on this topic: see for instance Davies 2001, Kivy 2002: 202-250, Dodd 2007, Levinson 1990 and 2011, Giombini 2017.

4 Cf. Bertinetto 2012a: 108-113.

5 I use here Puy’s 2018 (335) paraphrase of Bertinetto 2012a, which, with a pragmatic self-contradiction, confutes his own claim that the meaning of my expression “the type/token theory does not fit for musical improvisation” is unclear.

6 It is ironic that by denying this difference between improvisation and the performance of a composition, one actually makes an argument in favor of the thesis I defend, because this move sounds very much like admitting the paradigmatic character of improvisation for the musical practice. However, two clarifications are in order here to avoid misunderstandings (such as Puy’s 2018: 334-338). (1) As I have argued, there is an important difference between creativity in performing interpretation of musical works and creativity in improvised music. In a sense, both may be thought of as improvisation (which, as I argued, is a matter of degree), but only the second is intentionally improvisational, meaning that the aesthetic content of the musical production is generated intentionally on the spot. This implies that, if an interpretational performance generates its aesthetic content on the spot (which is a matter of degree, to be judged on a case by case basis), then it is an improvisation (as it happens in the majority of the interpretations of jazz tunes: see Bertinetto 2012a: 119-120; Bertinetto 2016: 142-160). (2) That said, and as I defend in this paper, the difference between interpretation and improvisation is completely compatible with the idea that improvisation is the paradigmatic musical practice in virtue (a) of the performative nature of music and (b) of the kind of “logic” – a retroactive normativity – that rules the interaction between musical works and performances. Unfortunately Puy 2018 misses this point.

7 Definitional properties are properties that are parts of the definition of the item and constitutional properties are properties having the power to give an organized existence to something, forming parts or constituents of something.

8 I have discussed them in Bertinetto 2012a, 2012b and 2016.

9 Of course, these are generally properties of events and, obviously, also performing interpretations of musical works possess them. The point (again misconceived buy Puy 2018) is that performing interpretations are events that realize a musical work (a construct that becomes concretely existent only through the performances), and, according to the type/token ontological theory, a normative-prescriptive relation is in place between the performances as tokens and the work as ideal type. Precisely this normative-prescriptive relation does not rule improvised performances: there certainly may be tokens of many types, but, as I have argued elsewhere (Bertinetto 2012a: 109-113; Bertinetto 2016: 118-129), either (a) they do not token the ideal or normative types that are the only types relevant for the type/token musical ontology (that must account for the possibility of correct and incorrect performances of a work) or (b), if they instantiate normative types (a musical genre for example), they do so in a way other than the one the structuralist version of the type/token theory allows for performances of musical works. However, my point is stronger: the structuralist type/token theory does not work for musical ontology, because it misconceives the working of normativity in the relation work/performance, as improvisation exemplarily shows.

10 Briefly, exaptation means making the situation or the environment appropriate through and for the performing action. It is complementary to adaptation to the situation or the environment.

11 I thank an anonymous referee for pushing me on this point.

12 Actually, a mental reading of a score may result in a determination of the musical work. As an anonymous referee suggested, this could be conceived of as a silent performance of the work. The problem is, however, that this performance cannot be perceived and evaluated: nobody can know whether it is a good or a bad one. People need real performances, not imaginative ones, for experiencing musical works and assessing their aesthetic and artistic value.

13 We may accommodate this idea within the type/token ontology. In order to do so we should however abandon Platonism (or structuralism) and conceive of types not as abstract, uncreatable and unchangeable platonic universals, but rather as plastic constructs that may historically change due to performing and cultural interpretations. Arguing in favor of this theoretical possibility is work for another paper, but I consider this article as a contribution that begins to take this path. I thank Enrico Terrone for this suggestion.

14 An open system develops its own limits (See Luhmann 1995) in an autopoietic way (see Maturana and Varela 1980). Fischer-Lichte 2004 famously applied the notion of autopoiesis to performance art.

15 For this “Gadamerian” view of the relationship between musical works’ performances see Feige 2014.

16 Indeed, as we have seen, normativity is improvisational, since the rules, to be valid, must be plastically adapted to the specific concrete situation, and this, in turn, involves the transformation of the norms “as we go along.”

17 Cf. Condello and Terrone 2017 (43): “As a historical outcome, a genre has an origin and it can undergo changes, which are triggered by new works that acquire exemplarity by virtue of negotiations within the genre itself as a practice”.

18 Cf. Condello and Terrone 2017. In my view (which is openly evaluative and not classificatory) this should be understood in an evaluative sense. I cannot elaborate on this point here, but I mean that artworks that are not able to impact on artistic genre in some way, being mere instantiations of artistic categories and not contributing in a qualitative way to the (trans)formation of categories are “less” artistic (thus, less artworks) then artworks that – also through the interpretations they generate – contribute to creatively (re)shape artistic categories.

19 For the notion of signifying in relation to improvisation see Monson 1996; cf. Bertinetto 2017b.

20 Thus, even granted Laetz’s interpretation of Walton’s paper (see Laetz 2010), according to which the relevant category a work belongs to is 1. perceptual and 2. “aesthetic active,” the point would be that the (sense and/or the effects and/or the efficacy of the) aesthetic activity of an artistic category may change historically (which happens with the evolution of taste, for instance), thereby impacting on artworks’ ontological identities. I thank A. Andrzejewski for bringing Laetz’s paper to my attention.

21 Notions like emergence and emergentism can be applied to the artistic field in different ways. For an explanation of different kinds of artistic emergentism and a defense of the philosophical productivity of this view see my Bertinetto 2019, in which I explore further aspects of the theoretical approach proposed here.

22 Cf. Bertinetto 2017a.

23 In Bertinetto 2017b I resorted to this idea for explaining musical meaning.

24 Cf. Hallam-Ingold 2007.

25 Previous versions of this paper have been presented in Paris on March 22nd 2018 (at the workshop L’expérience esthétique comme “praxis”: perception, imagination et atmosphères) and in Warsaw on November, 8th 2018. I thank all participants – and, in particular, T. Griffero, G. Bertram, A. Bondì, A. Andrzejewski and M. Salwa – for their comments and questions. I thank also two anonymous reviewers for their careful reading of the manuscript and their insightful suggestions.

Torna su

Per citare questo articolo

Notizia bibliografica

Alessandro Bertinetto, «Improvisation and ontology of art»Rivista di estetica, 73 | 2020, 10-29.

Notizia bibliografica digitale

Alessandro Bertinetto, «Improvisation and ontology of art»Rivista di estetica [Online], 73 | 2020, online dal 01 février 2021, consultato il 18 septembre 2021. URL:; DOI:

Torna su

Diritti d’autore

Licenza Creative Commons
Rivista di Estetica è distribuita con Licenza Creative Commons Attribuzione - Non commerciale - Non opere derivate 4.0 Internazionale.

Torna su
  • Logo Rosenberg & Sellier
  • DOAJ - Directory of Open Access Journals
  • OpenEdition Journals
Cerca su OpenEdition Search

Sarai reindirizzato su OpenEdition Search