Autonomous Morphology in Diachrony:

comparative evidence from the Romance languages

Faculty of Medieval and Modern Languages

Faculty of Linguistics, Philology & Phonetics

University of Oxford Logo

University of Oxford
















What is the Project about?


'Irregular verbs' are the bane of many a learner of Romance languages. It may surprise those who have struggled to memorize the irregular verbs of, say, French, Italian or Spanish that they are, in fact, fascinating entities, which occupy a fundamental place in the history of the Romance verb and pose major theoretical problems for linguists concerned with morphology and with language change. This project explores the evolution through time, and variation through space, of what could informally be described - and probably strikes most learners - as 'organized nonsense' in the verb. Taking the history of the inflectional paradigm of the verb across the Romance languages (including thousands of little-known dialects) we explore the origins, extent and nature of recurrent structural patterns in the verb-system which, despite lacking any apparent 'raison d'être', survive, spread and are reinforced through time (and how they die out, when they do). We suggest that such phenomena are far more pervasive and important determinants of morphological change than is traditionally assumed, and discuss why they are persistent over time. As an aid to this end we are producing the first ever comprehensive comparative database of the inflectional morphology of the Romance verb, the Oxford Online Database of Romance Verb Morphology, through which we investigate the neglected role of 'autonomously morphological' (or ‘morphomic’) structure in the history of inflectional paradigms. This should be of interest whether you are a theoretical morphologist, a Romance linguist, or simply curious about the (sometimes astonishing complex) details of verb-forms not only in well-known Romance languages but also in some very little known ones. The Database will be accessible from 1st April 2011 at



The other major outcome of the project is a volume entitled The Romance Verb. Autonomous Morphology in Paradigm Change, now under contract to Oxford University Press.


  Over the fours years for which it has been funded (October 2006 - December 2010) the Project has produced a full description of the extent, and nature, of the major and recurrent Romance morphomic patterns sketched in preliminary earlier work, particularly by Martin Maiden. This includes a greatly enriched typology of phenomena subject to morphomic distribution, beyond mere stem alternation. These include: defectiveness; alternations motivated by purely pragmatic factors; syncretism; the distribution of inflectional endings; conjugation class membership (heteroclisis). We also identify and describe other - minor and more localized - morphomic patterns. We also have new insights into the way morphomes can come into being and die out. For example, in addition to loss of phonological conditioning and loss of semantic conditioning of paradigmatic alternation, the differential effects of frequency in the spread of analogical changes has been found to be a source of novel morphomic patterning. On the other hand morphomic patterning tends to dissolve, not (as one might expect) as a result of any tendency to 'align' alternation patterns with extramorphological (e.g., semantic) conditioning, but usually as an accidental result of sound change. In other words, the most common cause of the genesis of morphomic alternation can also be one of the causes of its demise.


 Our findings also lead us to reject what might be described as the 'morphomic ghetto'. Earlier work on morphomes (notably by Aronoff and Maiden) tends to identify as 'morphomic' only  phenomena for which any extramorphological motivation can be ruled out. Our research has forced us to confront serious challenges to Maiden's earlier claims that certain alleged morphomic phenomena are not extramorphologically conditioned. Our conclusion, however, is not that the morphomic status of the relevant phenomena should be rejected, but rather that the boundary between the morphomic and the non-morphomic is often gradual and that, correspondingly, phenomena traditionally ascribed purely to phonological or semantic conditioning often contain an irreducibly morphological component. We have also found evidence to suggest that the degree of 'morphomehood' may be sensitive to the type of functional alternation at issue.


Our work has also demonstrated the importance for the study of Romance morphology, and general historical morphology, of two relatively neglected domains:


i. Romanian dialects, and particularly the remarkable body of morphological data contained in the (still only partly published) materials of the Romanian regional linguistic atlases.


ii. Non-European Romance varieties. Our original assumption (and one probably shared by most Romance linguists) was that the most interesting developments in verb morphology were likely to be limited to the European dialects, overseas varieties being colonial offshoots of the standard varieties (or of dialects similar to the standards). Our work on Acadian French as spoken in Canada has shown that this assumption is unsafe, and that theoretically significant developments have taken place in 'colonial' Romance varieties.


 A by-product of our research has been a bibliographical database of Romance (verb) morphology, containing some 5000 entries (including references to relevant works in theoretical morphology). This is not publicly accessible at the moment, but we may be able to make available some parts of the database on request. The bibliographical database also contains an increasing number of references to works on nominal and adjectival morphology.







Project funded by the
Arts and Humanities Research Council
(grant AH/D503396/1)
Site Disclaimer
Last update:  2 December 2010