The Representation of Initial Voiced Stops as Unvoiced Stops in Anglicised Brittonic Place-names.

Introduction

Watson analyses the name Prenbowgall (Watson, 1926: 351) as being composed of equivalents to Welsh pren "tree" and bugail "shepherd" and goes on to discuss the occurrence of the word pren in Scottish placenames. James, however notes that initial devoicing of stops is often seen in Brittonic names in the "Old North" (that is the historically Brittonic speaking of area of Northern England and Southern Scotland whose language has been called Cumbric) and that Prenbowgall might in fact be bryn bugail (James, 2012). Other names in the Old North with this feature of p- for b- are  Plenmeller  Blaenmoelfre (Ekwall, 1960: 368),  with blaen  as "plen", probably in Plenderleith near Kelso.  Paisley is said to come from Brittonic basaleg  We also have t- for d- in Tennis Castle, Drummelzier which is dinas  (Watson, 1926). 

However, the representation of Brittonic names with initial voiced stops as unvoiced stops in English is not restricted to the Old North.  If we look at other parts of Britain we find Tenby for Welsh  Dinbych in Pembrokeshire, Tintern for Dindyrn in Monmouthshire (Owen & Morgan, 2007) and Tintagell for Cornish Dintagell (Padel, 1988). In terms of personal names we also have Taffy for Dafydd.

An initial possibility might be that voiced initial stops were devoiced as a development of the Cumbric language. However, the fact that we see this feature in names adopted into English in Wales and Cornwall where we know that the initial voiced stops were certainly not devoiced in Welsh and Cornish, suggests that this is not a feature of the Cumbric language in fact but a phenomenon arising when English adopted Brittonic place-names.

We have noted how some names with initial voiced stops in Brittonic are rendered in English by unvoiced stops. The problems is that this is not uniform. If it were, we would expect to see the earliest borrowings such as Dover and Berkshire borrowed as *Tover and *Perkshire. So the earliest borrowings (though in fact their date may be a red herring and it might be their geographical location which is the determining factor, see below) and the later borrowing in Wales and Cornwall preserve the Brittonic voice.  We therefore need a solution which explains the rendering of Brittonic initial voiced stops as Old English unvoiced stops at certain times and places.

Brittonic vs Anglo-Saxon phones.

A phoneme is a segment of sound which, by contrast with other segments of sound, creates semantic differences. In English tip is a different word from dip (International Phonetic Association, 1999). The significant difference is the voice added to the initial stop in dip. Furthermore, phonemes are not identical across languages. For example in Indian dialects [v] and [w] are allophones of the Hindi phoneme /v/ but to English ears they are different phonemes namely /v/ and /w/. In Icelandic [k^h] and [k] are two different phonemes but in English they are one cf. kit [k^hɪt] but school [sku:l], where both seem to represent the same sound to English ears, namely the phoneme /k/, though they are phonetically different.

Phonemic contrast within a language does not seem to be an on/off phenomenon. It has been argued that rather it is a gradient (Scobbie & Stuart-Smith, 2006). In the production of any speech-sound there will be different elements such as aspiration, voice and length. Goblirsch notes how linguists have talked of fortis/lenis contrasts where the phonemic distinction is produced by producing a sound either strongly or weakly. He feels that talking of fortis/lenis is merely a over vague way of meaning contrasts in voice, length and aspiration (Goblirsch, 2010).  Aspiration here does not mean spirantization but merely the presence of a noticable breathing after the production of the stop. In specific languages, some of these features will be redundant in terms of producing a contrast, while others will produce phonemic distinction.

The addition of voice became a very important grammatical marker in Brittonic after the lenition of British consonants (Jackson, 1953)

We know that voice initially or in internal position was not phonemic for fricatives in Old English. Therefore the allophones θ ~ ð, f ~ v, s ~ z in internal positions were without any semantic inference and therefore were perceived by the Anglo-Saxons as the same sound. Laker argues that it was the adoption of Old English by historically Brittonic speakers who spoke a language, where θ and ð were different phonemes, that produced the later phonemicisation of these sounds in English (Laker, 2009). Modern English speakers perceive s and z for example as different sounds but Old English speakers did not and phonemicisation of initial fricatives is though to have occurred with French influence (Blake, 1992).

In contrast to fricatives where in Old English  voice contrast was not important initially, the generally accepted position is that the difference between initial voiced/unvoiced stops was phonemic in Old English (Hogg, 1992: 89).  However, it might not be voice that was the distinguisher, but aspiration. Goblirsch (2010) says that initial p,t,k in Southern Old English had developed distinctive aspiration before vowels while b,d,g initially had lost distinctive voice. The North Sea Germanic dialect closest to English - Frisian - shares the pattern of aspiration with Southern English. However not all English dialects had (or have) this pattern and Northern English and Scots (except the Highlands due to Gaelic influence there) do not have aspirated initial stops.

Goblirsch says that "In English… Gmc. p, t, k, have been aspirated initially before vowels and sonorants, except following s-, while NWGmc. b, d, g, have been fully or partially devoiced initially. Medially, b,d,g have remained voiced in a voiced environment…" (Goblirsch, 2010: 187) But he adds, "not all modern English dialects have aspiration… While there is aspiration in the South and Midlands there is generally no aspiration in Northern English or Scots." (ibid: 188)

Spaargaren (Spaargaren, 2007: 115) talks about different language groups, those which use voice to make phonemic distinction and those which use aspiration.  She says that English is an "aspiration language". She says that the distinction is between aspirated and unspecifed "plain" stops. This may be misperceived as a voice distinction where in fact it is an aspiration distinction.

Taken together with Goblirsch's comment that p,t,k were aspirated in Southern Old English and b,d,g were fully or partially devoiced, we have a phonemic opposition in Southern Old English between aspirated initial stops and "plain" (that is unaspirated) stops. In referring to German Spaargaren says that the “voiceless” stops in German can be written as follows:  Pass /p^h­as/   while unaspirated stops are written bin  /b̥ɪn/  Spaargaren characterises English with German as having a contrast between voiceless aspirated stops in foot-intial position pin  /p^hɪn/ and  attack /athhak/ and “plain voiceless obstruents” such as bin  /b̥ɪn/.  

Spanish stops are not aspirated and this leads a Spanish speaker to hear an English “b^h” more like a Spanish “p” than a Spanish “b”. This leads anecdotally to the famous Beatles song as heard in Spain "Let it Pee".  

So the difference for Spaargaren in present day English is not a voiced/unvoiced contrast but a plain/aspirated contrast. I presume she is using standard Southern English in view of what has already been said about Scots and Northern English not having initial aspirated stops. This system seems ancient in that Goblirsch seems to be saying it arose pre or immediately post English settlement for southern dialects of Old English. So when Old English speakers in Kent and Middlesex came across Brittonic names with aspirated initial stops they were able to accommodate them directly to their own aspirated stops. However, when Anglian settlers further north encountered Brittonic aspirated initial stops (which they did not have in their phonetic inventory in their northern English dialect) the foreign aspiration cued them to ascribe the initial voiced (but aspirated) Brittonic stops to their own unvoiced phonemes, because Brittonic  /bh/ sounded more like their /p/ than their / b̥/.

Laker cites Weinreich's ideas of an 'over-differentiation of phonemes'  where speakers of one language hear allophones of another and assimilate them as representing phonemic distinctions - as Weinreich puts it "involves the imposition of phonemic distinctions from the primary system on the sounds of the secondary system, where they are not required" Laker citing Weinreich, (1979: 18)

We actually see that to English speakers Welsh initial voiced stops still sounded like English unvoiced stops as late as Shakespeare's time. German (2009) notes that Shakespeare’s representation of the Welsh pronunciation of English shows that what should be initial voiced stops sounded unvoiced to English ears when pronounced by Welshmen.  in Shakespeare’s Fluellen Shakespeare’s representation of his Welsh accent has him pronouncing born as “porn” and brains as “prains”.  Shakespeare invites mockery by having Fluellen pronounce Alexander the Big (i.e. Great) as “Alexander the Pig” (German 2009).

Brittonic d- , -d-, -d

Old English had a d-, -d- but not a -d until Middle English. Brittonic had d, -d- and -d.

A quick look at Watson (Watson, 1926: 372) shows there are examples of intial Brittonic d- remaining as d- in Anglicised placenames such as in Dunpelder (Dinpeleidyr). As far as internal Brittonic internal -d- we see it both as -t- as in Tintern but also as -d- in Dunpelder. The -t- in Tintern can be explained as it is at the beginning of the second syllable, a privileged position which is treated almost like being word initial according to Jackson (Jackson, 1953).

Internally the Anglo Saxons seem to have borrowed Brittonic -d- as d,  We see the names such as Redmain rhyd maen and personal names such as Caedmon for Cadfan and Caedwalla for Cadwallon. However, we do find the Cumbric/Welsh personal name Gwaspadrig "St Patrick's Servant" as Cwaespatrick (Jackson, 1953: 390) showing not only the devoicing of the final [-g] in accordance with the general English system but also internal [d] being rendered as [t]. If Rutter Falls in Westmorland represents rhaeadr  "cataract" then internal [d] is also devoiced. These names seem to be borrowed much later than the earlier examples of Caedmon and Caedwalla and may be as late as 10th or 11th Century.

Finally, -d can exist in Old English cf tīd "time". However, sometimes final -d is rendered -t as in the Afon Tanad , River Tanat in Powys. In Cornish final stops become unvoiced in unstressed syllables. (Williams, 2012). This might be due to the influence of English as this does not happen generally in Breton or Welsh. Fulk says that a tendency to devoice final voiced stops in noticeable in the English of the West Midlands and in Scots (Fulk, 2012). It seems that any Brittonic -d appearing as English -t can be seen as a process within English not a Cumbric dialect feature.

Brittonic g-, -g-, -g

Brittonic g- arises from British/Latin g-. Medially and finally it arises from British -k- .

In the earliest Old English there was no initial /g-/. The sound was /γ/ initially and medially. The initial /γ/ became a stop /g/ by Classical Old English according to Hogg (Hogg, 1992) so therefore around 900. Borrowings from Brittonic after this can be adopted with g-.  Jackson discusses Förster's view that some early borrowings of Brittonic g- were rendered with k- but he dismisses it. Generally Brittonic names beginning with g- were substituted by early Old English γ which then was treated as Anglo-Saxon and palatalised as in Yarrow from garw and Yeavering from Geifrfryn.

Medially there was no -g- in Old English and it was regularly substituted with -k-.  This can explain for example Dacre apparently from Deigr and Melkinthorpe apparently containing the personal name Maelgwn and the river Cocker from a British *kukrā giving a Welsh *cogr.

Names such as Culgaith represent borrowings when internal -g- developed in early Middle English due to Norse or French influence, therefore 10th Century at the earliest.

In combination -tg- was borrowed as Old English "cc" /tʃ/ as in Litchfield in the 7th Century (Hogg, 1992; Jackson, 1953: 333)

Brittonic b-, -b-, -b

In Old English intial b- occurred freely but medial -b- is only in geminates -bb- (Hogg, 1992)

Then we have blaen names in Cumberland which generally remain b-  such as Blencow Blaencau, Blencogo Blaencwgw, etc.  Internally we see names such as Tenby where medial -b- is retained. 

References

Blake, N. F. (1992). The Cambridge History of the English Language: 1066-1476 (p. 732). Cambridge University Press. Retrieved from http://books.google.com/books?id=UlD3ksfXl5IC&pgis=1

Ekwall, E. (1960). The Concise Oxford Dictionary of English Place-names (4th ed.). Oxford: Oxford University Press.

Fulk, R. D. (2012). An Introduction to Middle English: Grammar and Texts (p. 500). Broadview Press. Retrieved from http://books.google.com/books?id=Wde9KnhS6UkC&pgis=1

Goblirsch, K. (2010). A Historical Typology of the English Obstruent System. Anglia - Zeitschrift für englische Philologie, 127(2), 176-207. doi:10.1515/angl.2009.026

Hogg, R. M. (1992). The Cambridge History of the English Language: The beginnings to 1066, Volume 1 (p. 633). Cambridge University Press. Retrieved from http://books.google.com/books?id=Y83EPUUWnQkC&pgis=1

International Phonetic Association. (1999). Handbook of the International Phonetic Association : a guide to the use of the International Phonetic Alphabet. Cambridge U.K. ;;New York NY: Cambridge University Press. Retrieved from http://www.worldcat.org/title/handbook-of-the-international-phonetic-association-a-guide-to-the-use-of-the-international-phonetic-alphabet/oclc/40305532&referer=brief_results

Jackson, K. (1953). Language and History in Early Britain. Edinburgh: University of Edinburgh.

James, A. (2012). Brittonic Language in the Old North. Scottish Placename Society. Retrieved July 16, 2012, from http://www.spns.org.uk/bliton/blurb.html

Laker, S. (2009). An explanation for the early phonemicisation of a voice contrast in English fricatives1. English Language and Linguistics, 13(02), 213. doi:10.1017/S1360674309002998

Owen, H. W., & Morgan, R. (2007). Dictionary of the Place-Names of Wales. Llandysul: Gomer.

Padel, O. J. (1988). A Popular Dictionary of Cornish Place-Names. Penzance: Alison Hodge.

Scobbie, J., & Stuart-Smith, J. (2006). Quasi-phonemic contrast and the fuzzy inventory: examples from Scottish English. Edinburgh: Queen Margaret University College.

Spaargaren, M. (2007). LARYNGEAL SPECIFICATION IN PRESENT-DAY AND HISTORICAL ENGLISH (pp. 112-128). Edinburgh: University of Edinburgh. Retrieved from http://www.ncl.ac.uk/linguistics/assets/documents/8.Marleen_Spaargaren-FT.pdf

Watson, W. (1926). Celtic Placenames of Scotland. Edinburgh.

Weinreich, U. (1979). Languages in Contact: Findings and Problems (p. 164). Mouton. Retrieved from http://www.amazon.co.uk/Languages-Contact-Findings-Uriel-Weinreich/dp/9027926891

Williams, N. (2012). Desky Kernowek: A Complete Guide to Cornish (p. 492). Evertype. Retrieved from http://www.amazon.co.uk/Desky-Kernowek-Complete-Guide-Cornish/dp/1904808999

Collins, B. and Mees, I. M. (2003) The Phonetics of English and Dutch. Brill.

Goblirsch, K. (2010) “A Historical Typology of the English Obstruent System” Anglia – Zeitschrift für englische Philologie, 127 (2) Available at http://nro-dd.sagepub.com/lp/de-gruyter/a-historical-typology-of-the-english-obstruent-system-UzNvjAxZZt

Kim, H., Kamiyama, T. and Halle, P. (2009)“Phonetic Enhancement in Feature-Driven Loanword Adaptation: A case of Japanese speakers’ perception of Korean stops”  Accessed [05.02.2011]Available at http://www.ling.ed.ac.uk/ocp6/ocp6-abbk.pdf

Prehn, M. (2009) “Fortis/Lenis in North Low Saxon” Accessed [05.02.2011]Available at http://www.ling.ed.ac.uk/ocp6/ocp6-abbk.pdf

Spaargaren, M. (2006) “Laryngeal Specification in Present-Day and Historical English”