Source
Descola (2016), LIT-822 (HAU 6(3):33–44, a keynote lecture), read in full in NOTE-641. Santucci, Doja and Capocchi (2020), LIT-807 (Symmetry 12(10):1706), read in full in NOTE-632.
What was actually shown
A structure is a set of transformations (LIT-822). Elements are characterized “not by intrinsic properties but by their position vis-à-vis one another”; the model “provides the syntax of the transformations which allows one to move from one variant to another”. “A structure is not a system”: it requires invariant relations that allow moving from one set to another by a transformation.
Two regimes (LIT-822; C2, C3 moderate).
- Goethean. Variants develop an initial form. In Les structures élémentaires, reciprocity develops from restricted exchange between two moieties to generalized exchange, to its combination with bride purchase among the Kachin. The variants are generated by a rule.
- Thompsonian. Continuous deformation between given forms, after D’Arcy Thompson, which Lévi-Strauss cites for myth. But unlike organisms, whose differences rest on the discontinuities of the genetic code, the group of a myth’s transformations is “a virtual continuum upon which the analyst performs arbitrary cuts” (citing L’homme nu, 603–4), so “each analyst will trace the path of transformation which his or her fancy will suggest”. D’Arcy Thompson needs stable forms at both ends; myth analysis has none.
The encoding (LIT-807). A myth is a list of (term, function) mythemes. Homology, inversion, opposition and symmetry are each “replace term a by term t in every mytheme”, with t supplied by the user; adding and removing a mytheme complete the set (§2.2). The canonical formula is explained by the Jealous Potter, where the key step (the ovenbird inverts the nightjar) is ethnographic knowledge, and no rule computes it. The Discussion concedes that the inverse of the ogre, and every such value, “is not obtained from the software system”, and that without it “it is impossible to define the transformation”. The validation generates 70 Mythologiques myths and 28 Corsican tales “by selecting each time a set of transformations to be applied on a given myth”.
What could have come out otherwise. An implementation could have derived the oppositions from the variants, or produced Lévi-Strauss’s orderings from the reference myth alone. LIT-807’s does neither, and says so.
What this does not say
- That Lévi-Strauss’s analyses are wrong or arbitrary in content. The claim is about where the path between variants comes from, not about whether a given analysis is apt. Descola names no criterion for preferring one ordering to another, and the record holds none.
- That every structural analysis is constructed. In the Goethean regime the variants follow from a rule, and Descola puts his own four ontologies there (C4, stated, not defended here).
- That the software tests anything about myth. Reproducing given analyses with operations chosen to reproduce them cannot fail (NOTE-632, C4). It is evidence here only through its concession about the table.
- That the method cannot be made non-arbitrary. Fixing stable end forms, as Thompson’s method needs, or inferring the table from a corpus, are open (NOTE-641 and NOTE-632, Open questions).
- That the distinction is Lévi-Strauss’s. The Goethe/Thompson names are Descola’s. Lévi-Strauss credited Thompson, and Goethe only “in passing”.
Connections
- Doja, Capocchi and Santucci (2021; LIT-804, NOTE-609) summarise the same engine faithfully and add a map display of one folktale. NOTE-609 reads it the same way: which change to make is supplied by the user. It adds nothing this account rests on and is not a source.
- CLAIM-019 holds that Lévi-Straussian transformational analysis, as Descola reads it and as computationally modelled, is prior art for structural transport. This account fixes what that prior art contains: an encoding of the analyst’s transformations, not a procedure that finds them. ARG-014 draws the same conclusion for trope annotations.
- Propp’s morphology (THEORY-167) is the syntagmatic contrast Dundes draws (LIT-830): an order of functions read off the text, against Lévi-Strauss’s paradigmatic regrouping. Descola does not mention Propp. NOTE-641 adds, as the reader’s own connection, that Propp’s Goethe epigraphs announce the Goethean regime.
- Lévi-Strauss’s Structural Anthropology (LIT-775) is held unread.