In film we have a story, a screenplay developed from the story and the film that is the screenplay realized. Few people would consider the screenplay a work unto itself other than perhaps film scholars. The director of the film gains the acknowledgement as the chief creator, and it is in the completed film that one is drawn to look for the "art". In composed music, there are scores and there are performances. A great composition is more likely to be considered a masterpiece than any particular performance, though most people would rather hear the music than read the score. This dual nature of the composed musical composition complicates the task of locating the core of the work.
A musical score is a relatively complete representation of a piece of music, and musical notation is, of course, a kind of language. Some people are able to read a musical score and imagine the musical results in their head, although this is a very specialized ability possessed almost exclusively by musicians and composers with a long history of hearing music. To anyone who does not have this ability, the music contained in the score is inaccessible. If we are thinking in terms of making music accessible to people who are deaf, it is not clear that a musical score would provide much of sense of the music. The score could certainly be analyzed, and a certain pleasure could be derived from that, particularly from a structural perspective, but this would still be a far from complete sense of the musical experience of the work. And of course, much music is never written out as a score.
Other music may have a score, but the score may give little indication of the experience that the composer intended even to a gifted score-reader. The composer James Tenney, for example, often wrote music where he explored the play of harmonics in the ear of the listener. His piece "Spectral Canon" does exists as a score (see pages 4 and 5 of this paper to see score examples), but the score is very unlikely to give the reader any sense of the intense play of harmonics that result in the sweeping waves of shifting timbres that make this piece a astonishing listen. This piece is built in a very mathematical way, and math is certainly not modality specific, so there are many ways to represent the pattern of this work, but it is unlikely that many would carry the visceral excitement of the sound. See his related work Spectral Variations for an audio example.
Some composers eschew any form of traditional notation, creating graphic scores that give a sense of the desired musical timbral colour, dynamics, shape, and flow. These scores are often used by composers whose ideas do not fit comfortably within the expressive scope of traditional musical notation. They are inherently open to subjective interpretation, there being no standard and agreed upon ways to interpret any of the graphic gestures on the score. See "Pictures of Music" to browse a large selection of graphical scores. Unlike conventional musical notation where the composer must find a way to express his or her musical intention within the limits of the notation, graphic notation allows the composer to invent ways of visually expressing the most significant aspects of the composition. These scores make this music inherently more accessible to the those unable to hear the music.
The music comes to life in the hands of musicians. Part of the musicians' job is to make the music hidden in the score accessible to the general hearing public. The scope of the interpreter's task varies widely even within music where this pairing of score and interpreter is prevalent. Some composers seek a mechanical reproduction of the score and are rigidly opposed to any subjective modulations on the part of the interpreter. Others leave significant details of the actual performance open to the interpreter and declare that the work is not complete without the interpreter's contributions. The interpreter often invests a composition with a significant portion of its musicality and emotional impact, breathing nuance, based partly on established traditions of expressive embellishment and partly on personal taste. The most celebrated musical interpreters are "stars" in their own right. In jazz and most improvised music, the role of interpreter is very often more significant than the role of composer, and in free jazz, the composer and score were dispensed with altogether.
Composers like John Cage created scores which essentially functioned as machines which produced the actual score and/or the performance based on the results of chance operations like rolling dice or flipping coins. Cage was seeking ways of impeding himself from actively influencing the production of the work. In other cases, the instructions themselves are intentionally ambiguous and left to the performers to interpret. It is part of the difficulty of locating the "work" in the case of Cage's indeterminate pieces that Cage, despite his desire to distance himself, still had very clear ideas of what was a proper realization of these works and what was not. In Cage's case, it is often not clear that the experience of a particular performance is important, but also not clear that the instructions or the indeterminate score suffice in any way to represent the work. Some of his works (HPSCHD) were indeed explicitly experiential, verging at times on the psychedelic.
Underneath this question of the score versus performed music or concept versus object is a fundamental concern of this inquiry: To what degree can you treat any given work of arts as an object of cognition, something to be processed by the brain directly, rather than as an integrated experience of sense / perception / cognition, which necessarily presents greater accessibility challenges for those with sensory impairments?