by Craig Evans
The Bristol Centre for Linguistics’ seminar series continued yesterday with an engaging talk from recent PhD graduate Harry Parkin. The subject of Dr Parkin’s presentation was what tax returns reveal about the West Midlands English dialect; or to be precise, what the poll tax returns of 1377, ’78 and ’81 reveal about specific dialectic features. The main feature under examination was the rounding of /ɑ/ before nasal consonants, which would become an <o>. To illustrate this point, Dr Parkin provided the etymological example of band and bond, which would have once denoted variant spellings of the same word.
One of the most intriguing aspects of Dr Parkin’s research, and diachronic linguistics in general, is the methodology involved. How can we really know how people spoke in a time before technological innovations enabled voices to be recorded? The answer: well, we can’t really know for sure. However, evidence for the sound of English in the past does exist: in the form of the written word.
At a time before standardisation had taken root, the spelling of many words will have been influenced by how they were pronounced in different dialects. This is especially true of the spelling of names, which naturally would reflect the speech of the individual holders of those names. For Dr Parkin, then, the names included on poll tax returns provided a rich source of data. For the purposes of looking at the prenasal /ɑ/ and /o/ variation in West Midlands English, the names Dr Parkin focused on for his study were those that included /man/ or /mon/.
As well as the advantage of providing data that is less likely to conform to orthographic standards, the poll tax returns are also very useful for revealing the geographical distribution of dialect. Previous research has often relied on literary sources to pinpoint on a map how different dialectic features are distributed. The problem with this approach is mainly one of provenance, such as uncertainty about dates and the geographical origin of the scribe. Also, as Dr Parkin points out, language variation is part of a continuum where dialect and geographical boundaries are not clear cut, so a more representative data source is needed.
In addition to the prenasal /ɑ/ and /o/ variation, Dr Parkin also considered the occurrence of <u> before -s, -l, and -r in unstressed syllables. While the former reveals a high concentration of /mon/ in some of the counties looked at to indicate a West Midlands dialectic feature, the same couldn’t be said of the latter which was found to be too widely distributed.
In the final analysis, the main question that arises from this study, one that Dr Parkin identifies himself, is: how much can we trust that the spelling of the written word represents the phonology of the spoken word?
It is possible that the scribes who wrote the names on the tax returns may have used their own dialect preference when spelling those names. When we look at the tax return data, therefore, are we seeing the distribution of dialect or just the dialect of the individual scribe? While this may never be known for certain, Dr Parkin contends that there would have been a natural incentive in tax collecting to keep spellings familiar to the different localities. Perhaps a topic for future research will be to investigate the identities and origins of the scribes involved in producing the poll tax returns – after all, a point that was raised during the post-talk questions mentioned how the dialect of scribes may be useful as an indicator of dialect distribution.
The next talk in the series will be on 12 November: ‘Some More Aspects of the Pragmatic Theory of Properhood’ by Richard Coates.