Optional theme: Aesthetics - generative art, authorship and mutability
The Harbourlight Gallery has budgeted £18,000 to acquire 'Tidal Archive', a generative installation by artist Mira Okafor. The work uses an algorithm trained on 4,000 images of local fishing boats; each time it is displayed, it produces a new 12-minute projection and a unique soundtrack. Okafor supplies a seed value and a 40-page 'score' but does not control the output. The gallery's funding contract requires acquisition of an 'original work of art by a named artist'. A local collective claims the work is collectively authored because the training images include their family photographs, for which they gave no consent. Okafor argues that authorship lies in the algorithm and score. The acquisitions committee must decide whether to buy, how to label authorship, and whether the work's mutability undermines its status as a single artwork.
1. Assess whether 'Tidal Archive' is a single artwork or a series of distinct works.
Read the worked answer
Distinction: a type is the repeatable work; a token is a particular display. Goodman's autographic/allographic distinction helps: an autographic work is identical with its physical instance; an allographic work is specified by a notation and can have multiple correct instances. Okafor's score, seed and algorithm function like a notation, though it underdetermines output.
Application: each 12-minute projection is a token; the work is the type defined by score + algorithm + seed. Mutability does not by itself destroy unity - many works are variable, such as mobile, performance and software art. But if the algorithm's outputs are not constrained enough to be recognisable as this work, the type is too thin. The 40-page score and fixed training set give enough constraint.
Verdict: treat it as one variable-media work with multiple instantiations, not a series of unrelated works. Label: 'variable media installation; each display is an authorised instantiation.' This matters for funding: 'original' can mean the type is original, not that every token is unique.
2. Evaluate Okafor's claim to authorship against the collective's claim.
Read the worked answer
Okafor's case: authorship is located in the conceiving act - selecting the training set, writing the score, choosing the seed. This is a strong intentionalist claim: the artist sets the conditions of display. The algorithm is a tool, not a co-author.
Collective's case: the training images are family photographs used without consent. Authorship is not the same as source contribution, but moral rights and consent are separate issues. If the work's meaning depends on those images, the collective has a claim to acknowledgement and possibly to credit as a source community, not necessarily joint authorship. Anti-intentionalists would say the work's meaning exceeds Okafor's intention, which strengthens the collective's interpretive stake.
Test: ask what would be lost if Okafor were sole author. The score and algorithm would remain, but the ethical provenance would be hidden.
Verdict: Okafor is the artist-author on the score and algorithm; the collective deserves a named provenance credit and a consent agreement. Joint authorship would require shared control over the work's conception, which they do not have. The committee should not let authorship labelling erase the consent problem.
3. Recommend acquisition terms.
Read the worked answer
Recommendation: acquire for £18,000 only with a revised contract.
Terms: (1) credit 'Mira Okafor, artist-conceiver' and 'Source community: the local collective - training images used with/without consent'; (2) require Okafor to deposit the score, seed values and a documented training-set manifest; (3) obtain retrospective consent or remove identifiable family photographs; (4) label the work 'variable media' in the funding report, explaining that 'original' refers to the conceived work, not a fixed object; (5) include a community consultation clause before first display.
Justification: the funding contract's 'original work by a named artist' can be satisfied by an allographic variable work; originality is not the same as immutability. But the consent gap is a separate ethical and legal risk. If the collective refuses consent and the images are central, the committee should either renegotiate the price or decline. This protects aesthetic integrity and provenance.