Exaptation from Arabic syntax to Persian lexical Morphology

It has long been natural for linguists to invoke metaphors taken from the life sciences, in particular evolutionary biology. In recent decades it has been argued that such analogies represent not mere rhetoric or metalinguistic convenience, but actual processes in language development; after all, why should not a biologically-conditioned cultural phenomenon such as language, the collective product of a life form, replicate aspects of a form of life – and not just general processes such as evolution and extinction, but actual details of the process, such as natural selection and exaptation? “Exaptation” is a biological term coined by Stephen J. Gould and Elisabeth Vrba in a 1982 article. It may be defined as “a case where an anatomical structure that originally evolved to serve one function was later commandeered to facilitate a quite different function.” For instance, feathers in the proto-avian dinosaur lineage demonstrably evolved before the capacity for flight; their purpose must have been something else, such as thermo-regulation (to keep them warm), or display (to attract a mate) – functions which additionally continued in many cases. However, this structure was fortuitously available when it was later modified to provide flight-control surfaces. Darwin anticipated this process in 1859, applying the term “pre-adaptation,” and citing the example of a fish’s swim-bladder, as having originally evolved for flotation, and in land animals being converted to a wholly different purpose – that of respiration, in the form of lungs. In linguistics the term has been adopted by, e.g., Roger Lass in a 1990 article, and by Laura A. Janda in Back from the brink: a study of how relic forms in languages serve as source material for analogical extension (Lincom Europa, 1996). My example of the redeployment of a distinctive feature to a novel purpose (which I investigated in the1980s, before I had heard of exaptation) involves the successful hijacking, by semantic determinants of Persian lexical morphology, of a syntactically-conditioned phonological alternation in the feminine ending of Arabic nominals. The human motivation was the desire, during the development of the Eastern Islamic Kulturgebiet between the seventh and twelfth centuries (see the map, fig. 4) to incorporate useful or prestigious Arabic vocabulary into Persian, using the Arabic writing system (which had been adopted in Persian) but adhering to Persian phonotactics and lexical morphology. The junk element in Arabic feminine-ending nouns and adjectives was not so much that Persian had no grammatical gender, but that it had no use for the typically Semitic syntactic structure known as the “construct state”: this requires that a feminine ending be pronounced as /-at/ with terminal -t when its nominal is the head of a NP modified by a following noun (“pre-juncture position”), and /-a/ in all other situations (“prepausal position”). Thus dawlat al-Sūdān ‘the state of Sudan’, but ra’īs al-dawla ‘head of state’ and dawla mustaqilla ‘an independent state’. In each case, the feminine marker was written with an invariant hybrid graph in Arabic. Now, Persian speakers needed definitively to lexicalize a single form of a word as either -at or -a. The solution they devised has led to an inventory of at least 1400 Arabic Feminine Ending (AFE) loanwords in the modern Persian lexicon, in a ratio of roughly 600 -at: 800 -eh, including forty doublets – i.e., copies of the same word in each form, written with distinctive graphs (final t, and final h for spoken /-a/) and two lexically distinct meanings (c. fig. 3; I use -eh to represent the vocalic termination as being visually quite distinct from -at, and to