HENRYK ELZENBERG AS A FORERUNNER OF ANGLO-AMERICAN CONCEPTS OF EXPRESSION

Classic expression theory identified the emotional content of works of art with the feelings of the artists and the recipients. This content thus appeared to be external to the work itself. Consequently, formalism declared it to be irrelevant to a work’s value. A way out of this predicament – one which the Polish aesthetician Henryk Elzenberg (1887–1967) was among the first to propose – was suggested by the idea that physical, sensory objects can themselves possess emotional qualities. Thanks to Bouwsma and Beardsley, this concept – of expressiveness as a quality – became common in Anglo-American aesthetics from the 1950s onwards. At the same time, these authors demanded that the term ‘expression’ be expunged from the language of aesthetics. But the widespread tendency to conceptualize the emotional content of art in terms of the expression of a certain subject (most often the artist) still requires some explanation – interpretation, rather than negation. One interpretation construes the expressiveness of works of art in terms of the expression of a fictitious subject, the ‘work’s persona’, conceived by Elzenberg in the 1950s and 1960s. This article discusses his concept and explains some of its more complex aspects, before addressing the emergence of a very similar concept within Anglo-American aesthetics. This concept was gradually elaborated in the 1970s and 1980s, but only in the 1990s did it become more fully developed and widely discussed.


I. INTRODUCTION
The nineteenth century made its mark on aesthetics by leaving it under the influence of the paradigm of expression. This paradigm was first made credible in artistic practice by Romanticism. Then, in the late nineteenth and early twentieth centuries, it was given a theoretical formulation by figures such as Eugène Véron in  Art, 1938). 1 Despite all the differences between these systems, their common denominator seems relatively obvious: the fundamental content of a work of art is the creator's feeling, which is expressed in his or her work, allowing the recipient to share it. Such an understanding of expressiveness in relation to art has unquestionable persuasive power: it makes art something greater than a superficial aesthetic pleasure, something with important, profound content, a medium of spiritual understanding between people.
In this vision, however, a work of art is merely a medium for transmitting feelings from the artist to the recipient, and in this instrumental role it may be perceived as secondary and replaceable: any other object or method providing the same transmission of feelings is equivalent to a work of art. This stands in contradiction to another very strong belief about art -namely, that each work of art is autonomous and has unique value.
In other words, classic expression theory reduces the content of art to things that are essentially external to it, to the experiences of the artists or the recipients or both. In an attempt to restore value to the work itself, formalism deemed external elements of this type to be irrelevant to the work's value. In this way, interpreting the emotional content of art as comprising the experiences of the artist or the recipient results in taking emotion out of art. This conclusion is difficult to accept, since the emotional content of art is universally perceived to be an integral part of it (and in certain forms of art, such as music or lyric poetry, crucial).

II. EMOTIONAL COLOURING ACCORDING TO HENRYK ELZENBERG…
A way out of this predicament was suggested by the idea that objects perceivable by the senses (let's call them 'sensory objects') can themselves have certain emotional qualities. The Polish aesthetician Henryk Elzenberg  was one of the first to propose this concept, in 1937, at a time when later versions of classic expression theory were being formulated in the English-language literature (for example, Curt Ducasse, 1929, andR. G. Collingwood, 1938). 2 In 'Emotional Colouring as an Aesthetic Phenomenon ' (1937) Elzenberg drew attention to a way of understanding expressivity, 3 which is linked directly with the expressive object itself, for example, a work of art, and tends not to be overtly present in traditional theories of expression. 4 In order to explain and justify his conception, Elzenberg began by enumerating three phenomena that were traditionally encompassed by the notion of expressivity. The first of these is the expressing of real mental content by means of objects accessible to sensory cognition (for example, an artist's manifesting his or her experience in a work) and the second is the arousing of emotional states in the recipient; classic expression theory has come to focus on these two aspects. As the third phenomenon, Elzenberg enumerates 'animization' (animizacja), that is, ascribing a fictitious psyche to inanimate objects. 5 Thus we can say, for instance, that the sea 'grows angry' or that a weeping willow 'has grown sad' . This phenomenon is also considered by Elzenberg to be generally perceived in contemporary aesthetic literature.
But, over and above these three, he draws attention to a further phenomenon, one that he claims is more recondite than the others: a 'pure, subjectless emotional "quality" residing in an object […] that […] we might also figuratively call […] the emotional "colouring" of an object' . 6 By giving 3 No distinction between expressivity and expression in art is being made yet; the appropriate distinctions follow shortly. Elzenberg, 'Emotional Colouring' , 213. Elzenberg himself (ibid.) says that he was inspired by the suggestions of three authors: the psychologist Oswald Külpe (1862Külpe ( -1915 and the phenomenologists Max Scheler  and Roman Ingarden (1893Ingarden ( -1970, though he does not indicate which texts he has in mind. In a later article, he gives a reference only to Külpe, Grundlagen der Ästhetik, ed. Siegfried Behn (Leipzig: Hirzel, 1921), 101-2. See Henryk Elzenberg, 'Non-aesthetic and Aesthetic Expression' , in this issue of Estetika, 224n5. As for Ingarden, the idea appears, for example, in the The Work of Music and the Problem of Its Identity, trans. Adam Czerniawski (Berkeley: University of California Press, 1986). In this treatise Ingarden argues: 'We have […] to distinguish the emotional qualities rooted in the sounding material of the musical sound-constructs from feelings that the listener might experience [and from the] expressed feelings […] of the performer […] or of the composer. ' (p. 101) Although the Polish original of this treatise did not appear until 1958, the signature beneath its conclusion ('Paris 1928-Lviv 1933-Kraków 1957 shows that it was a long time in the making and that work was certainly further prolonged by the impossibility of publication during the Second World War and then the Stalinist period in Poland (see n. 42 below). But, before publishing of Elzenberg's article in 1937, he may well have been aware of Ingarden's article 'Zagadnienie tożsamości dzieła muzycznego' [The problem of the identity of a work of music], Przegląd Filozoficzny 36 (1933): 320-62, on which the later treatise was based. This earlier article already mentions subjectless emotional qualities which can also belong to certain non-mental objects (pp. 359-60). examples of expressivity which cannot be reduced to the three earlier types, Elzenberg substantiates his argument that such a phenomenon does indeed occur. When, for instance, someone says that a landscape is 'gloomy' or 'cheerful' , this cannot mean that the real feelings of some animate being are manifesting themselves in it, nor does it necessarily mean that the observer, under the sway of the landscape, is overcome by such a feeling. As Elzenberg says, 'For lyric poetry, the divergence between our own mood and the mood of the world around us was such a common theme that it became rather banal ' . 7 There remains animization: the notional ascribing of a 'gloomy' or 'cheerful' psyche to a landscape. But when we use emotional terms in relation to objects not endowed with a psyche, are we always dealing with animization, with attributing a fictitious psyche to objects? When we perceive a sea as 'angry' , do we only do so when we are inclined to say that it is 'angered' or that a weeping willow is 'sad' when it is 'saddened'? Apparently not, and Elzenberg also declares himself 'in favour of the independent existence of emotional "colouring" and against the possibility of reducing it to any sort of lower-level animization' . He argues that 'all animization seems to presuppose [the prior existence of emotional "colouring"] and cannot occur at all until the observer has perceived some emotional colouring' . 8 This is because animization is never neutral and does not involve ascribing to an object a qualitatively indefinite psyche: 'We do not animize "in general"; a fort or a mountain peak is animized specifically as "proud", spring as "joyous", a sea on a stormy day as "angry", and so on.' 9 It is only because we discern in an object certain qualities which we perceive as emotional that we are at all inclined to animize that object, that is, to ascribe to it a fictitious psyche which might underlie the qualities we have already discerned. So the perceiving of these qualities must be the prior phenomenon.
As we shall see below, it is precisely this phenomenon that Elzenberg considers the basis and essence of all aesthetic types of expressiveness, and thus of expression in art in particular.
But the assertion that the phenomenon of emotional colouring exists raises a question. Why do we describe certain qualities of inanimate objects using terms relating to feelings that such objects obviously cannot have? Elzenberg provides an answer, though with the reservation that it is only a suggestion and not a proposition of whose truth he is fully convinced, in contrast to the question of whether emotional colouring actually exists: Ibid.
[I]f we feel compelled to use an emotional term to define that colouring quality, it is perhaps because, at a given moment, some emotional fact does actually occur and assumes a particular relationship with the object we are beholding; and the only place where this fact could arise is in the psyche of the observer himself. 10 However, was emotional colouring not supposed to be a quality present in the sensory object itself, whose perception can be independent of the observer experiencing a corresponding feeling? Elzenberg suggests that it is possible to avoid a contradiction between the last two statements by postulating that the emotional fact discussed in the foregoing quotation need not be fully developed emotion, such as exists in our real mental life: And here is the hypothesis: that emotional fact would be emotional only through a certain quality that it possesses; besides that, it would discharge a cognitive function […]. It would be utterly devoid of such features of actual emotion as motivational force or a tendency to spill over into the whole of one's awareness and take over the whole of its content; its only function would be to reflect in the mind a certain quality of the object beheld -that very quality which we are calling emotional colouring. 11 It does not appear inconsistent with Elzenberg's intentions if we say that this fact was supposed to be the act of experiencing a pure emotional quality outside the wider spectrum of mental phenomena with which real-life emotion is typically associated or of which full-blown emotion usually consists: most frequently it has a certain cause, usually a cause which we are aware of, typically some object or situation (we are happy or sad about something). As long as this cause lasts, it is difficult for us to free ourselves from the emotion it evokes. The emotion (love, anger, envy) is frequently directed towards something: it has an object. It is accompanied by some idea or expectation. Lastly, it can entail a certain inclination or desire to act. But the reaction evoked by the perception of the emotional colouring of a sensory object is devoid of such elements: it is limited to the very act of experiencing a pure emotional quality. Since this emotional quality is not directed towards any object, 12 it does not give rise to any expectations or any conscious motivation to perform specific actions. And since the only cause of this experience is perception of a sensory object, we retain complete control over our experience of this emotional quality: it disappears as soon as we decide to stop beholding that object. 10 Ibid., 216.

11
Ibid. 12 Our attention is directed to the appearance of the object we perceive, but the emotional quality recognized in it is not. If someone were to claim that music is expressive of love, anger, or jealousy, then obviously not of love for, anger towards, or jealousy of the piece of music.
Elzenberg's suggestion may be seen as anticipating a theory of expression in art, which was later called 'weak arousal theory' . 13 (In that case, Elzenberg's vision would certainly be the weakest possible version of such a theory.) It may also be seen as a way of reconciling such theory with the so-called cognitivist position, which is perceived as being opposed to it. 14 The classic (strong) version of arousal theory states that expressiveness simply amounts to the capacity to evoke certain emotions; that is, that something is expressive of a certain emotion if and only if it evokes (or at least has a tendency to evoke) that emotion in recipients. The reasons for rejecting this simplistic view have been presented many times; almost no one attempts to defend it.
Elzenberg's remark (quoted above) on the divergence between our own mood and the mood of the world around us could legitimately be used as a standard argument against it.
The opposing view, sometimes called cognitivist, states that we do not experience the emotions expressed in a work of art; we merely recognize them as qualities of the work. On this view, expressive qualities are ascribed to the work of art itself and not, as in arousal theory, to elements external to it -namely, the feelings of the recipient. This might seem to be the position of Elzenberg as well, who, as we noted, sees the essence of the aesthetic variety of expression in the perception of emotional colouring.
Many authors express dissatisfaction, however, because this position appears to remove emotional reaction completely from the description of our interaction with art. Do we not value art precisely for its ability to move us? Would we be inclined to describe it in emotional terms if it did not arouse in us the emotions ascribed to it? In connection with this, these authors attempt to restore at least a limited role to aroused emotions in accounting for expressiveness in art, even if they do not go so far as to equate expressiveness with the evocation of emotions in a simple, naïve way, as was done above. That is why these theories are called 'weak arousal theories' or, more frequently, 'emotivist theories' . Since Elzenberg speaks about an emotional fact taking place in the recipient in connection with his perception of certain sensory objects (for example, works of art), prompting us to use emotional terms to describe their qualities, he may be cautiously deemed a representative of arousal theory. At the same time, this is perhaps the weakest possible version of such a theory. Not only is there no mention of full-blown, real emotion, but the emotional fact mentioned here is of the most attenuated kind, and its function is cognitive. We experience no full, real emotion, but only draw on our emotional sensitivity to recognize a certain quality of the observed object and define it as emotional colouring. In this way, Elzenberg could also be deemed an adherent of the cognitivist position. The only point on which he diverges from the orthodox cognitivism of Kivy is his acknowledgement that our recognition of emotional qualities involves not merely understanding them or reading them, but experiencing them or feeling them, which appears to correspond better to our experience of interacting with art and is a welcome concession to the emotivist position.
This type of compromise between the two positions appears close to Alan Goldman's formulation with respect to music: The correct middle view is not Kivy's new position that music has a tendency, unactualized for many listeners like him, to arouse ordinary emotions. It is rather that emotion states that are not ordinary or paradigm occur in the full engagement of typical listeners […]. The unusual features of these emotions […] are explained […] by our experiencing them in the context of being engaged in other ways as well with the works to which we react. Full appreciation […] normally requires some attention to form, for example, as well as affective reaction, and attention to such other matters is sufficient to block full-blown emotional reactions. 15 The diagnosis that our reaction to expressive qualities is not ordinary, full-blown emotion that is expressed is absolutely correct and immediately renders superfluous the arguments of the cognitivists, who most often strive to prove just this. 16 Unfortunately, the conclusion that they typically draw from this correct and uncontroversial proposition is that if we do not experience the emotions expressed by music in a full-blown, ordinary way, we do not experience them at all. But if for some reason, for example, because of shadow or twilight, we do not perceive the colours of some object, and thus do not see it completely, that does 15 Goldman, 'Emotions in Music' , 66. 16 See, for example, Kivy, Music Alone, 158ff. To be sure, Kivy grants that music may be deeply moving but he asserts that the emotions aroused are not those which are expressed. The object of the emotions aroused is music itself which 'moves us by various aspects of its musical beauty or perfection' (p. 161), not by its expressive qualities. Consequently, the emotions aroused do not explain, according to Kivy, the expressiveness of music.

Krzysztof Guczalski
Estetika: The Central European Journal of Aesthetics, XLIX/V, 2012, No. 2, 00-00 not imply that we do not see it at all: we might, for instance, discern just the outline of its shapes. And just because we might not recognize the instrument with which some melody is being played (and perhaps we cannot even unambiguously make it out if, say, our radio has poor quality speakers), that does not mean we do not hear the melody itself. Similarly, it seems entirely credible that in addition to ordinary, full-blown emotions we can, under certain conditions, feel certain abstract, attenuated versions of them in the type of emotional fact described by Elzenberg.
However, while making the correct diagnosis, Goldman gives a rather unconvincing argument in favour of it: 'Full appreciation […] normally requires some attention to form, for example, as well as affective reaction, and attention to such other matters is sufficient to block full-blown emotional reactions. ' 17 But what if someone does not attend to form and other non-expressive qualities of a work of art? Does he or she then experience 'full-blown' emotion? It appears that the emotional reaction to music is almost never of this type, since it does not include the typical elements of full-blown, ordinary emotion discussed above (the sources and causes of emotion, its objects, expectations connected with it, ideas or desires to act). This is not the subject's own emotion really experienced by it, but rather an external emotional quality that is merely beheld, similarly to how a moviegoer beholds the action presented in a film without actually participating in it. In other words, a full-blown emotional reaction does not have to be blocked, because listening to music simply does not give any basis for such a reaction.
We may also approach Goldman's argument from the opposite perspective, by considering ordinary emotion in real life: in experiencing such emotion, we are also sometimes engaged in other ways with the objects and situations that give rise to it: we look at a list of grades and are happy about a good grade in an examination, but at the same time we notice that our surname has been misspelled; we feel uneasy when travelling on the underground at night alone with just a couple of suspicious-looking types for company, and yet we punch our ticket, answer a question posed by one of our fellow passengers about the next stop, and so on. But does this mean that our full-blown emotional reaction is blocked by our attention to those additional matters, that we are not experiencing ordinary emotion anymore but rather the type of emotion that accompanies listening to music? That would obviously be a very strange conclusion.
It appears that our development of Elzenberg's suggestion better describes phenomenon of emotional colouring. Even if we accept Elzenberg's suggestion that in response to the perception of some sensory objects a minimal emotional fact with a cognitive function occurs and explains our tendency to describe these objects in emotional terms, another question arises: Why do these kinds of emotional facts occur with respect to some, but by no means all, sensory objects'?
To argue that it is because they have emotional qualities would obviously be circular. What is presumably called for is an answer in terms of more fundamental, elementary qualities traditionally ascribed to sensory objects, qualities that can be ascribed to them without controversy. This matter will be addressed in the second part of the following section.

III. …AND IN AMERICAN AESTHETICS
In American aesthetics, one of the first steps in the direction of seeking the expressivity of a work of art in the work itself -and not in the experiences of the artist or the recipient -was taken by Susanne Langer (1895Langer ( -1985. In Philosophy in a New Key, Langer sharply criticizes the view that music's expressivity is to be explained in terms of the expression of real feelings, experiences, or other emotional states in the composer or the performer (which she calls selfexpression) and that its emotional content consists in arousing certain feelings in the recipient. 18 The emotional content of a work of music is contained in the work itself, and Langer explains the grounds for this presence as follows: The tonal structures we call 'music' bear a close logical similarity to the forms of human feeling -forms of growth and of attenuation, flowing and stowing, conflict and resolution, speed, arrest, terrific excitement, calm, or subtle activation and dreamy lapses -not joy and sorrow perhaps, but the poignancy of either and both -the greatness and brevity and eternal passing of everything vitally felt. Such is the pattern, or logical form, of sentience; and the pattern of music is that same form worked out in pure, measured sound and silence. Music is a tonal analogue of emotive life. 19 But the aestheticians usually associated with a fundamental breakthrough in thinking about the expressivity of art are Oets Kolk Bouwsma  and Monroe Beardsley (1915Beardsley ( -1985. In his essay 'The Expression Theory of Art' , from 1950, Bouwsma arrives at the conclusion that the typical models which were normally used to explain the phenomenon of expressivity in art -namely, the down and lead to misunderstandings. 20 He therefore proposes abandoning those analogies and accepting that what we usually call the expressivity of works of art amounts simply to certain properties belonging to those works themselves: now, unabashed, we shall say that the music is sad, and we shall not go on to say that this means that the music expresses sadness. For the sadness is to the music rather like the redness to the apple, than it is like the burp to the cider. And above all we shall not, having heard the music or read the poem, ask, 'What does it express?' 21 A similar stance was adopted by the much better known and influential American aesthetician Monroe Beardsley, in 1958: The Expression Theory has called our attention to an important fact about musicnamely, that it has human regional qualities [elsewhere Beardsley also speaks of feeling qualities]. But in performing this service it has rendered itself obsolete. We now have no further use for it. Indeed we are much better off without it. 'The music is joyous' is plain and can be defended. 'The music expresses joy' adds nothing except unnecessary and unanswerable questions. For 'express' is properly a relational term; it requires an X that does the expressing and a Y that is expressed, and X and Y must be distinct. When we say that a rose is red, we have only one thing, namely the rose, and we describe its quality; in exactly the same way, when we say the music is joyous, we have only one thing, namely the music, and we describe its quality. There is no need for the term 'express' . 22 But why do we use emotional categories at all to denominate these properties? Bouwsma's answer to this question is by no means original (and Beardsley's opinion in this matter is similar): 'Sad music has some of the characteristics of people who are sad. It will be slow, not tripping: it will be low, not tinkling. People who are sad move more slowly, and when they speak, they speak softly and low. ' 23 Henryk Elzenberg as a Forerunner of Anglo-American Concepts of Expression In this way, he subscribes to one of the two main traditions occurring over the course of history. Like so many things in philosophy, they originate in the writings of Plato and Aristotle respectively. According to the first, which could be called the paradigm of externality, 24 music has emotional content because of the resemblance of its melodic, dynamic, rhythmical, and other structures to the typical forms taken by the natural expression of emotion in the human voice, body movement, and posture. According to the second, which in turn might be called the paradigm of internality, music has such a content by means of its expressive difference between the arts is that painting conveys feeling through a visual image, and literature through a rationally intelligible statement, but music conveys the naked feeling direct. ' 29 But by the middle of the twentieth century, it seemed time for another paradigm shift; that is, a return to the Platonic view of externality, as exemplified in Bouwsma's statement. Even though this conception subsequently dominated aesthetics during the second half of the twentieth century, 30 it still had its opponents, such as Malcolm Budd, who claims that its 'application is exceedingly narrow' and that 'music can penetrate beneath the surface of emotion to its innermost core: music is not restricted to the outer world of expression of emotion but reaches as far as the inner world of emotion itself. I believe that this is so. ' 31 Also Jerrold Levinson, who essentially subscribes to the externality view, at the same time proposes to reclaim 'the relevance that the dynamic and phenomenological dimension of an emotion's inner aspect, and not just its behaviourally constituted aspect, can have' . 32 And Aaron Ridley, who in 1995 accepted the prevailing Platonic paradigm, 33 in his later book discusses some arguments in favour of both the externalist and internalist positions. 34 Yet the two paradigms, the Platonic and the Aristotelian, of externality and of internality, are not as remote from each other as one might suppose, and they are certainly not in direct contradiction to one another. It is by no means implausible, and it is even to be expected, that the natural, public expression of certain feelings bears structural and dynamic similarities to our privately felt experience of those feelings as we perceive them introspectively. Hence one might suppose that the two theories actually converge with respect to those feelings which have natural forms of expression associated with them. Yet the two theories are certainly not entirely equivalent.
Only the externalist explanation appears to be empirically verifiable, because it refers to the intersubjectively accessible, external expression of emotions.
Undoubtedly, it was this advantage that allowed it to dominate analytic philosophy, most frequently with an empirical, naturalistic orientation -sober and straightforward. Kivy certainly finds this to be the essential advantage of this   But the connection between music and the external expression of emotion appears to be relatively superficial and seems not to explain why the recognition of this similarity should determine the value of music. On the other hand, the postulated similarity to the internal aspect of feeling appears to give music a more intimate connection with the essence of our mental life and enable deeper insight into it, thereby providing a more convincing explanation of music's resultant value. Moreover, internalism appears to have an advantage in that it can explain the presence in music of moods and feeling tones that are not typically connected with any external expression.
The leading advocates of either view enumerated above -from Plato to Kivyusually have not even mentioned the possibility of the opposite standpoint, let alone provided any arguments against it and in favour of their own. Thus it might seem that the choice of one or the other was down to the spirit of the times and philosophical taste rather than the conclusion of any argument. Only the most recent authors (Budd, Levinson, Ridley) take notice of both views. The main reason Budd gives for his preference for the internalist position is his strong belief, as expressed in the quotation above. Levinson merely suggests that the internalist view should also be taken into account besides the dominating externalist view (see the quotation above). Only Ridley presents some reasons to prefer one view or the other but without any hint that they might decide the case. In fact he even says: 'It may be, of course, that the truth lies somewhere between these views. ' 36 One might cautiously say that both positions are too one-sided to be able to refer to all types of emotional qualities in music or other sensory objects. Perhaps we should concur with Budd: 'a new theory of music is needed; and if this theory is to be revealing it will, I believe, have to be less monolithic than the theories I have rejected' . 37 So perhaps it is naive to think that it is possible completely to support one paradigm or the other to the exclusion of the other.
At least the externalist position is known to Elzenberg, who mentions one more possible explanation (in addition to the two discussed above) of the presence of emotional qualities in physical objects. 38 He does not, however, support any of the accounts. Likewise, he suspends his judgement on this matter in his later, longer, article on expression. 39 He was clearly of the opinion that all potential explanations were far from satisfying. Elzenberg could have refrained from 35 Kivy, Corded Shell, 68. 36 Ridley, Philosophy of Music, 74. 37 Budd, Music and the Emotions, 176. passing judgement on the question because, as we will see, the further part of his theory of expression is not dependent on any particular explanation of emotional colouring. We need only accept the fact of its existence, that is, that certain sensory objects have certain emotional qualities.
This assumption, accepted by Bouwsma and Beardsley, prompted them to make some quite controversial claims. As we recall, in interpreting the emotional content of a work of art as certain qualities of it, they simultaneously rejected out of hand any appeal to the human expression of emotion and distanced themselves from the language of expression in aesthetics. 'The Expression Theory […] has rendered itself obsolete.
[…] There is no need for the term "express". ' 40 '[W]e shall not, having heard the music or read the poem, ask, "What does it express?"' 41 This is particularly odd in the context of Bouwsma's explanation of the reasons why we are inclined to call music 'sad' (and Beardsley's opinion in this matter is similar), which points precisely to a similarity with the natural expression of emotions. They should therefore admit that even if 'sad' music is not simply an expression of someone's emotions, it is at least something 'expression-like' or quasi-expressive.
Moreover, even if we agree that emotional content in art is above all a question of certain qualities of the work itself, the widespread tendency to perceive and understand them as the expression of some subject (most commonly the artist) calls for some explanation (interpretation rather than negation) -an interpretation which at the same time would not (unlike some versions of classic expression theory) violate the equally common conviction of the autonomy and inherent value of the works of art themselves. Works of art are not, after all, reducible to merely the means of acquainting ourselves with the mental content of the artists.

IV. A RETURN TO EXPRESSIVITY AS EXPRESSION
This intuition was followed once again by Elzenberg, who in a later article, 'Nonaesthetic and Aesthetic Expression' , written in 1950, 42 formulates his initial problem Henryk Elzenberg as a Forerunner of Anglo-American Concepts of Expression argumentation leads him to a condition that expression must fulfil in order to be considered aesthetic: 'an image of mental content must be given not by the intermediary of a symptom, but together with it, directly, such that image and symptom might be grasped together in a single act of perception' . 44 And to illustrate his idea he makes a vivid comparison: mental content must be given in the symptom -or on it -more or less like moisture in a sponge, a scent in the air, sheen on snow or poetry in a sonnet, or, perhaps more bluntly, like wetness in water or greenness on a leaf. […] This content, to put it slightly less vividly, must be simply read by the observer from the symptom. Or to put it completely drily: the observer must find it on the symptom. In short, this may be called the immanence of the mental object in the symptom […]. 45 The parallel between this formulation and Bouwsma's 'redness of an apple' or Beardsley's 'redness of a rose' is, of course, patent.
So the mental content conveyed in aesthetic expression must consist of those mental qualities of the expressive object which are familiar from Elzenberg's earlier article. Therefore, whereas in the earlier article Elzenberg argued that something like emotional colouring belonging to objects that are perceptible to the senses does exist, he is now arguing that it is essentially the heart of expression in its aesthetic variety, in other words, it is the necessary condition of expression having aesthetic quality. Elzenberg agrees that the first, most fundamental, and, as he puts it, 'aesthetically purest' variety of the perception of emotional content in art is the perception of certain emotional qualities in a work of art. He observes, however, that quite often our perception does not remain exclusively on this level of purely qualitative reception: once a mental quality has appeared on an object, it insistently demands that some psyche -one in which it can 'settle' -should be produced, composed, or dreamt up for it. And […] before the observer knows it, the qualities have already drawn that psyche with them automatically, as it were, deep into the object. What follows […] is what we this time denote by the term animization [animizacja]: the attribution of a psyche to things not endowed with one. Of course, that psyche is not neutral, indeterminate; the object receives those experiences and dispositions which correspond to the quality observed [in the object]. 46 This is how we animize many natural objects in particular: 'mountains and rivers, wind and the night' . And the psyche ascribed to them is, for obvious reasons, 'entirely, unequivocally fictitious' . 47 However, Elzenberg also points to a further type of reception, a further variety of our attitude to objects characterized by emotional qualities, which is like the next stage in animization and concerns only such objects as are somehow associated with an actual human psyche -namely, human artefacts. Objects of this kind, particularly works of art, may of course be animized in the same way as natural objects: 'it is [not] Beethoven who at a given moment relinquishes his internal struggles and falls into an ecstasy of joy, […] it is the Ninth Symphony, the content-laden sound mass' . 48 In this case, there are two psyches associated with the object. First, there is the fictitious, imagined psyche of the Ninth Symphony, filled with exactly the content that we discerned as emotional colouring in the object itself. This is therefore the 'immanent psyche of an aesthetic object' . And then there is the real psyche of the composer, transcending the aesthetic object, and linked to it only genetically. This real psyche does not fictitiously belong to the aesthetic object, but actually belongs to some real human being -here, Beethoven. In this situation, there occurs, however, a natural tendency to identify these two psyches with one another: we now somehow identify the psyche that we ascribed by way of animization to the object with the psyche of an actual human being, with which the object maintains a factual link; we lose the sense of difference so thoroughly that we no longer have two psyches, the fictitious and the real, within our field of vision, but only one, which belongs to both the object and a person. It soon turns out, however, that this is not identification on equal terms: the real human psyche is stronger than the fictitious, flimsy mental life of the object and consequently has a tendency to oust it from awareness. And ultimately what was identification can even become the substitution of the human psyche for the psyche of the symptom: now, it is no longer the actual symphony that rejoices or breaks down, but that person within it, embodied in it, as it were, who thus shaped it. And the same applies in other cases: it is not through the symptom, but in the symptom itself, as its content, that the person manifests himself. 49 One might go so far as to surmise that these two psyches often do not even appear as separate in the awareness of the recipient, but rather from the start as a composite, which Elzenberg defines as the outcome of the process of identification, that is, as a conglomerate of the two: as a psyche which we call by the artist's name (and which we imagine to be his or her psyche) but which possesses qualitative endowments that are wholly derived from the emotional colouring of the work itself.
Thus the work, in a tautological way, becomes an adequate image and expression of that psyche -just as in ordinary animization. Unlike in ordinary animization, however, the expressed psyche is understood as the real psyche of the composer and not as the unequivocally fictitious psyche of the work. It is not the Ninth Symphony that 'grows angry' , but 'Beethoven' embodied in it.
Since the mental content which is manifest in the work belongs, at least notionally, to a certain being who is indeed endowed with a psyche (that is, the composer), and not, as in ordinary animization, to an object which is not actually endowed with a psyche and only 'possesses' one in an unequivocally fictitious way, we may reasonably conclude that we are dealing with an instance of expression in the proper sense. But because this content is at the same time directly present in the work of art as its emotional qualities, we may regard this expression as aesthetic: 'And that is just how expression in the proper sense, in its aesthetic variety, would look. Or, perhaps more in keeping with the actual state of affairs, that is what that expression, in order to be aesthetic, must transform itself into. ' 50 This means that it is not, literally speaking, expression in the proper sense. In truth, the mental content manifest in this expression does not belong to a real psyche; we only imagine that it does. In other words, we ascribe the mental content derived from the work itself to a certain imagined human subject, whom we most often imaginatively identify with the work's creator. Thus the 49 Ibid., 230. 50 Ibid., 231. perception of a work as the expression of its creator is only an elaborate rhetorical figure, even if we are sometimes not fully aware of its figurative character. It would therefore seem that in this case we understand expressivity as the expression of a fictitious human subject.

V. CONCLUSION
Elzenberg's conception appears to do justice to the strong and generalized tendency to understand art as expression -something which cannot be said of the contemporary conceptions of Bouwsma and Beardsley, which would banish expression quite one-sidedly from the language of aesthetics. At the same time, Elzenberg's interpretation does not undermine another powerful conviction, that of the autonomy of the work of art (after all, the 'artist's psyche' is derived entirely from the work) -an accusation that could have been levelled at classic expression theory from the late nineteenth and early twentieth centuries.
Unfortunately, Elzenberg's work was not translated into English. Anglo-Saxon aesthetics found itself under the overwhelming influence of the one-sided stance of Bouwsma and Beardsley. When this position eventually triggered opposition, aimed at rehabilitating the widespread use of the term 'expression' in relation to art, Anglo-Saxon aestheticians spent the next forty years or so elaborating a multidimensional and well-balanced conception similar to that devised by Elzenberg, at the same time naturally discussing many specific matters not raised by him. On the other hand, there also appeared attempts to rehabilitate the term 'expression' in its original, proper sense. One of the first such attempts, referring solely to the fine arts, dates from 1965. 52 In relation to music (although primarily music with words), a conception of the expression of a fictitious subject was first Henryk Elzenberg as a Forerunner of Anglo-American Concepts of Expression proposed, as Jerrold Levinson states, 53 by Edward T. Cone in 1974. 54 In 1982, music in general, not necessarily with words, was considered by Donald Callen, who notes that expressivity may be understood not only as the presence of some or other qualities of a certain impersonal object (an acoustic product), but also as the fictitious expression of someone's emotional states. 55 We find similar suggestions made in 1985 by Jenefer Robinson, in relation to literature, and in 1986 by Bruce Vermazen, in relation to art in general. 56 Not until the 1990s did the concept of the expression of a fictitious subject become more widely disseminated and discussed. The most mature, comprehensive, and persuasive version of this concept -in essence the closest to Elzenberg's, though more elaborate than his -is the theory put forward in 1995 by Aaron Ridley, again formulated in relation to music. 57 Like Elzenberg, Ridley sees the construct of a fictitious subject (called a 'persona' in this conception) as only one of the possible ways of perceiving music. 58 But in the following year (1996), another work appeared, Levinson's 'Musical Expressiveness' , burdened -as in Bouwsma and Beardsley -with the typical overstatement of a single model. Levinson's work includes the suggestion that the construction (notion) of a fictitious subject appears whenever we hear music as expressive, that the expressivity of music is simply (always!) hearing it as the expression of a fictitious subject. 59 So this is the opposite pole to Bouwsma and Beardsley. In turn, dissatisfaction with such a one-dimensional view has helped to inspire recent articles by Robert Stecker 60 and Saam Trivedi, 61 who state, correctly, that when receiving music as expressive, we by no means always imagine a fictitious person as the subject of that expression. Seeking an alternative conception, Trivedi makes the 'discovery' that another, frequent, way of perceiving the expressivity of music is … the animization of the music itself.
'Discovery' in inverted commas, since animization, as we remember from 53