In defence of morphomic analyses

ABSTRACT There has been some debate over the notion of “morphomes”, i.e., patterns of inflection without any clear motivation outside of morphology. Morphomes are taken as evidence for some autonomy of morphology. However, it has been claimed that there is “very little evidence for change which operates on morphology alone”, in other words that morphology does not change independently – and this has, consequently, been used as an argument against the “morphomic” approach. This paper presents evidence of inflection classes arising or being “strengthened” in Scandinavian, classes that do not have any function outside of morphology. This, then, is evidence of change, operating on morphology alone – even in the relatively “poor” inflection systems of Scandinavian. One of the case studies also shows affixes being changed in order to align better with non-affixal inflection. This goes against the claim that non-affixal inflection is epiphenomenal. The paper also presents criticism of some other arguments that have been used against the morphomic approach. Notably, the “diagnostic problem” suggested for morphomes is hardly more severe than that involved in many other approaches to morphology. The paper also shows a (perhaps unexpected) convergence between the morphomic approach and strands of functionalism. While morphomic patterns may seem redundant and local, this is not unique to them. Many generalisations done by language users may seem redundant and local to linguists.