Why do some pieces of music, or film, or visual art strike me as “great works of art,” independent of how much I enjoy them? When I introspect on this, and when I test my brain’s mostly subconsious “artistic greatness function” against a variety of test cases, it seems to me that the following features play a major role in my artistic greatness judgments:

1. Innovation
2. Cohesion
3. Execution
4. Emotionality

Below is a brief sketch of what I mean by each of these. In the future, I’ll elaborate on various points, and run additional thought experiments as I try to work toward reflective equilibrium for my aesthetic judgments.

Innovation

Rite of Spring. Citizen Kane. Young Ladies of Avignon. If someone produced a ballet, film, or painting very similar to these today, critics would not be impressed, and neither would I. But these works were innovative at the time, and pushing on the boundaries of an art form — in structure, subject matter, tone, concept, etc. — feels artistically impressive to me, all else equal. ¹

Among innovations, I think I’m most impressed by complex, non-obvious innovations. Beethoven’s addition of voices to the symphony was innovative, but it’s a pretty obvious idea and if he hadn’t done it, I suspect somebody else would have done it not long thereafter. In contrast, Carlo Gesualdo’s chromatic but highly emotional madrigals seem to have been innovative in a less obvious way, since nobody else wrote in such a chromatic style until centuries later, ² and then only in very different musical languages. Had Gesualdo not written emotionally extreme chromatic madrigals, it’s possible that nobody would have ever thought to do so.

Meanwhile, my brain doesn’t seem to associate influence with artistic greatness. (To illustrate: Gesualdo was innovative but he was not influential. Depeche Mode were influential but not innovative.) To me, a work of art’s influence seems much less a fact about the work itself than it is a fact about external factors, though influence surely provides some evidence about e.g. innovativeness and emotionality. Influence’s lack of direct impact on my judgments of artistic greatness might constitute one of the largest differences between what I intuitively think of as artistic greatness and what many others intuitively think of as artistic greatness.

Cohesion

In my meaning, an art work is “cohesive” if many or most of the artist’s choices in composing the work are united in serving a particular artistic purpose. ³ For example, a film in which the writing, acting, editing, and cinematography all seem to be specifically engineered to achieve a particular effect seems more impressive to me than a film where the editing, acting, and cinematography seem to be chosen “off the shelf” and only the writing is being used to achieve the film’s artistic intent. This is one reason I usually don’t feel artistically impressed by film adaptations of well-regarded plays, even when they are exceptionally well-written.

Consider The Matrix‘s innovation of bullet time. The effect is now used in many kinds of films, but I think it has special artistic merit in a film about people in a computer simulation, where someone able to control the simulation could presumably “freeze” the activity of the simulation and view its events from various angles. In this way, a very specific cinematographic choice coheres with the film’s subject matter in an interesting way.

Execution

Many art works seem innovative and cohesive to me, and yet poorly executed in some way. Max Richter’s The Blue Notebooks isn’t very innovative, but it seems to me basically flawless in execution, like a Bach fugue. Amogh Symphony’s Vectorscan is fairly innovative and artistically cohesive — in this case, by thoroughly embracing its ambition to smash together a huge number of musical genres (which in a different sense, I know, could be called anti-cohesive) — but the album seems sloppy and poorly executed compared to, say, Vampire Rodents’ similarly eclectic Clockseed.

It’s difficult to say fully general things about my execution criterion, because it’s the “devils in the details” criterion. Talking about why Karma is better executed than Free Jazz will have little in common with talking about why Mozart’s Jupiter Symphony is better executed than his Symphony No. 11.

Emotionality

Finally, my brain’s artistic greatness function seems to think that (human) art is greatest when it can be innovative, cohesive, well-executed, and also engage the human being emotionally.

These days, it can be a challenge to be both innovative and emotionally engaging — to a wide audience, at least. I don’t know much about evolutionary aesthetics, ⁴ but I would guess that humans are pre-programmed to be most emotionally engaged by a certain range of aesthetic stimuli (prior to lots of experience or training, at least), just like we are pre-programmed in so many others ways. By now, I would guess that many of the musical forms that are emotionally engaging to most people around the world with little experience or training — e.g. the tonal 4-minute song in 4/4 time — have been “artistically exhausted” in the sense that there aren’t many ideas compatible with that form that haven’t already been explored. Meanwhile, Royal Trux’s Twin Infinitives is pretty innovative, but relatively few people can stand listening to it for more than 60 seconds.

Of course, what one experiences as emotionally engaging can change over time as a result of experience. ⁵ There was a time when Klaus Schulze’s Irrlicht would have bored me to death, but now it sometimes brings tears to my eyes. Meanwhile, most of Anthony Braxton’s work fails to engage me, but I can imagine that it might engage me at some point in the future, in part because other people somehow do find it emotionally engaging.

Innate individual variation probably also plays a role in aesthetic preferences, but (I would guess) only within a certain range. ⁶ It’s certainly possible I will never be able to enjoy Anthony Braxton’s Trillium R the way some others do, no matter what training and experience I acquire, and perhaps even no matter what life history I could have lived. But I doubt those who like it now would have liked it had they heard it as infants, even if they were innately capable of learning to like it later.

I do suspect, however, that a vast range of sound patterns can be emotionally engaging given a certain history of human experiences. Maybe even John Cage’s HPSCHD would be emotionally engaging to someone who was exposed to similar randomly-generated music in the womb and as an infant.

Given all this, I think my brain awards “emotionality points” to works of art on the basis of a guess about how little experience, especially how little unusual experience, is required for humans to be emotionally engaged by the art work.

So, if someone managed to write an innovative, cohesive, well-executed tonal song in 4/4 time that even a feral child would instinctively bounce their head to, that would win a lot of artistic greatness points from me. In contrast, John Cage’s HPSCHD is in some ways innovative and cohesive, but my brain doesn’t award it any points for “emotionality.” Twin Infinitives gets some points for emotionality, Irrlicht gets even more, and Beethoven’s Heroic Symphony gets still more.

Footnotes:

1. Interestingly, some aspects of innovation can be investigated via large-scale data analysis. Here is an example from Dean Simonton, explaining some of his historiometric inquires in Simonton (1998):

   [I conducted a] computer analysis of all of the melodies that are likely to be heard in the classical repertoire. All told, I gathered data on 15,618 themes by 479 composers (Simonton, 1980b). Because a mere 100 composers account for 99% of all pieces frequently recorded and performed (Moles, 1958/1968), this was an inclusive sample of thematic material.

   The next step was to obtain the first six notes of each theme, and then transpose these notes to a C tonic… The notes were then grouped into five two-note transitions, and then the frequencies calculated by a computer (Simonton, 1984b, 1994). These frequencies provided the basis for estimating the repertoire melodic originality of each theme. A theme that scored low on originality contained extremely commonplace two-note transitions (e.g. dominant-dominant and dominant-tonlc), whereas a theme that scored high on originality contamed extremely rare two-note transitions (e.g. chromatic notes and unusual intervals).

   At this point one might wonder whether these computer-generated scores have any validity whatsoever. After all, how can a comp-uter judge the originality of melodies that it cannot hear? Yet these scores have been validated several ways. First, the scores based on twonote transition probabilities have been shown to yleld the same results as scores based on three-note transition probabilities (Simonton, 1980a). Second, the scores computed by the computer correlate positively with subjective (human) ratings a of melody’s arousal potential (Martindale & Uemura, 1983). Third, these scores are associated with pertinent aspects of a composition, including the key, the rhythm, the form, and the medium. For example, minor-key themes have higher melodic originality than major-key themes, and themes from chamber music display higher melodic originality than those from orchestral music (Simonton, 1987)…

   …Most creations in classical music contain many themes. This is especially true in the case of multiple-movement forms, such as the symphony, concerto, quartet, trio, and sonata. Therefore, I was interested in learning how the composer altered the magnitude of melodic originality as the piece progressed. The investigation concentrated on the large works of Beethoven, and the results were very interesting (Simonton, 1987). Melodic originality was not randomly distributed across the course of a composition, but rather it fluctuated according to a backward-J curve. The most originality appears in the first movement, the second most in the last movement, and the least in the middle movement(s). Evidently, the composer begins with themes that have some shock value. After attracting attention, originality can relax, as so often happens in the middle movements of the symphony — the slow song-form movement and the following minuet or scherzo and trio. Then the intensity picks up in the last movement, albeit without the need to reach the same heights as the beginning. This is probably unnecessary because of the contrast effect imposed by the intervening movements.

   …melodic originality has changed from the Renaissance to the 20th century (Simonton, 1980b). First, there exists an overall upward trend toward ever more original themes. Second, superimposed over this trend are cyclical fluctuations. In particular, melodies in the Renaissance were highly predictable, but by the time of Monteverdi and Gesualdo, originality had reached a maximum point. After that, themes became more predictable again, albeit not as predictable as in the Renaissance. The trough occurred during the classical period – Haydn, Mozart, and Beethoven. With the advent of the romantic period melodic originality increased, most notably with the chromaticism of Chopin, Brahms, and Wagner. This trend reached a new height about the time of World War I. It was this peak that witnessed the emergence of the atonal and serial music of Arnold Schoenberg. Then melodic originality saw another decline, although again not going as low as the two previous lows.

   …The ultimate aesthetic question is what determines the impact of a particular work of art. Therefore, I conducted a series of studies that examined whether melodic originality gauged by a computerized content analysis had any predictive validity by this criterion. I specifically used two criteria, subjective and objective.

   [Subjective ratings] were taken from Halsey (1976), who rated thousands of classical compositions on two dimensions, namely aesthetic significance and listener accessibility. Melodic originality was correlated with both of these measures, albeit in opposite directions (Simonton, 1986a). Aesthetically significant compositions are those that are deemed sufficiently profound that they can withstand repeated listening. They cannot be fully appreciated in a single, superficial hearing. Such pieces score high on the melodic originality of their thematic material. Accessible compositions, in contrast, can be understood at once, and thereby lend themselves to music appreciation courses taught at schools and colleges. Not surprisingly, works that score high on listener accessibility tend to score low on melodic originality. Highly unpredictable themes would only interfere with making an immediately favorable impression.

   [Objective ratings] were based on the actual success of the composition in the classical repertoire (Simonton, l98Ob, 1983b). That is, these ratings gauged the frequency that a musical product is likely to be heard in the concert hall, opera house, and recording studio. The relation between this objective measure and melodic originality is much more complex than holds for the subjective measures. The function is neither linear positive nor linear negative but rather curvilinear. In particular, the relationship is best described as an inverted backward-J curve (Simonton, 1980b; see also, Simonton, 1987). The most popular works are those that have moderate levels of melodic originality, whereas the most unpopular works are those with the highest levels of melodic originality — the pieces with low melodic originality having more middling popularity…

   For an overview of objective and subjective measures of aesthetic success used in the literature, see chapter 5 of Tinio & Smith (2014).[↩]

2. Except for a few immediate Gesualdo imitators.[↩]
3. Cohesion could naturally be considered a sub-component of execution, but it seems to play an important enough role in my judgments of artistic greatness that I’ve decided to separate it from other features of good artistic execution.[↩]
4. I would guess the evidence in this field isn’t very compelling yet, anyway.[↩]
5. The effect of experience on musical preferences may begin as early as week 18 of fetal development, when the Mayo Clinic claims the ears form and the fetus may begin to hear. There’s also a study (Soley & Hannon 2010), which I haven’t read, which claims:

   We presented 4- to 8-month-old American and Turkish infants with contrasting melodies to determine whether cultural background would influence their preferences for musical meter. In Experiment 1, American infants preferred Western over Balkan meter, whereas Turkish infants, who were familiar with both Western and Balkan meters, exhibited no preference. Experiments 2 and 3 presented infants with either a Western or Balkan meter paired with an arbitrary rhythm with complex ratios not common to any musical culture. Both Turkish and American infants preferred Western and Balkan meter to an arbitrary meter.

   Of course, those differences could be innate, but I doubt it.[↩]

6. Brown & Jordania (2013) make the case for musical universals, but I haven’t read their paper.[↩]