The third Old-World Conference in Phonology (OCP3) in Budapest has just finished. The programme had a number of very high-quality talks in a variety of different phonological frameworks, and the atmosphere was very good.

I will not give an overview of all talks — here is the conference website, including all the abstracts, but I felt that one could see two opposing trends in this OCP. On the one hand, there were some very good talks which were strongly grounded in a phonetic or psycholinguistic tradition, although these were not as many as there would have been perhaps a few years ago; instead, there were also several talks which argued for a much more abstract phonological point of view. Some took the radical step of adopting Element Theory or going back to Containment models of faithfulness (instead of Correspondence), but one could also hear arguments in favour of autosegmentalism, feature geometry, etc.

There seems to be a tendency to take representational issues more seriously, after a period of time in which constraint rankings alone were supposed to do all the work. In his famous book on phonology in the twentieth century, Stephen Anderson described this history as going back-and-forth between derivational concerns and representational concerns. With this he predicted the coming of OT after Autosegmental and Metrical Phonology, since OT is mainly a theory about the way derivations work — but now the pendulum seems to swing back again. Actually, also the attention to phonetic detail can be seen as a renewed interest in ‘representational’ issues: a spectrogram is also a representation.

Particulary enlightening for me was an interesting discussion between two of the keynote speakers of this OCP, Tobias Scheer (Nice) and Donca Steriade (MIT). Scheer made a division between ‘representationalists’ and ‘non-representationalists’ in his talk, and had put Steriade in the latter category. In the question period, Steriade pointed out — rightly, I think — that it is not possible to formulate constraints without referring to representations. However, it cannot be denied that concerns about autonomous properties of representations, those that are not to be understood directly by their grounding in the phonetics, has been weaker in the past ten years than they had been before.

Of course, there were also talks which were neutral in this debate, or dealt with completely different points of discussion. For instance, there were two very good talks on the theory of OT acquisition, and a number of talks about the phonology-morphology interface.

OCP4 will take place in January 2007, on the beautiful island of Rhodes (Greece).