Pagan Anglo-Saxons, Phonology, and Christian lexis 11

Miscellaneous speculative matters appended to the foregoing posts.

In the preceding ten posts I put forward the following theory:

In southern Britain in the fifth century there was a significant Latin-speaking Christian British population. This is evidenced by pre-550 Latin loanwords in Old English and in particular by Christian terminology. The loanwords are most likely to be due to this population adopting Old English as their first language. This may be why there is also an important group of religious terms coined in Old English rather than borrowed from Latin. And it also explains why there are virtually no British loanwords in Old English.

1. hlaford and hlæfdige

Assuming that these two compounds were formed at the same time, or at least that hlaford is unlikely to be more recent than hlæfdige, we can tell that they must be very old formations within Old English. Hlæfdige shows i-umlaut, which makes it pretty early if this umlaut is to be dated in the first quarter of the 6th century. But to have qualified for umlaut, the word needs to have undergone accelerated sound changes beforehand. Practically no lexical compounds exist in Old English in which a second element containing /i/ has been sufficiently reduced in stress or prominence to cause i-umlaut in the vowel of the first element (Campbell §204 (2)). There are apparently a few proper names and a bunch of grammatical or semi-grammatical adjectives and adverbs in which this has occurred. The nearest parallels to hlæfdige are endlufon (< *aːnliv-, undoubtedly a Germanic compound) and enwintre (< *aːn-wintr-; early enough for /æː/ to be shortened to /æ/) (both share in the raising of /æ/ from i-umlaut before nasals). So in hlæfdige the second element must have been worn down and been brought into direct contact with the root syllable of the first very early indeed. So early is the compound that the second element went out of use before written records began (compare the first element of weofod).

Similarly *hlaf-weard, the putative original compound underlying hlaford, is scarcely recorded, unlike other weard compounds (there is one occurrence, in the Paris Psalter: which may be an etymologizing guess, or even a re-coinage of the compound). It was already in its worn-down form before written records.

What is most likely to have caused this more rapid phonological attrition in both words is, one might surmise, vocative use. The words must have been in frequent everyday use from very early times. They were the everyday words for ‘lord’ and ‘lady’ in contrast to those we find in poetic diction, frea, drihten, and so on.

On the other hand, we know that ‘early’ doesn’t mean West Germanic. There is no trace of these words in the related languages, not even Frisian. So the formations occurred at a time when the OE speech community was no longer in contact with its close Continental relatives, most likely in the early stages of settling in Britain.

If these were the everyday words for the people in authority, what kind of society would have coined them? One where the people in authority were the guardian or master of bread and the kneader or baker of bread. These express agriculture-related rather than military-related roles. They also suggest a society in which raising corn is the key activity rather than herding animals.

One other compound of hlaf may be relevant here: Lammas. OED: ‘The 1st of August (Festum Sancti Petri ad Vincula in the Roman calendar; see also gule n.2), in the early English church observed as a harvest festival, at which loaves of bread were consecrated, made from the first ripe corn.’ There seems to be no connection between this festival and the synchronous Roman church festival of St Peter ad Vincula. No other nation has this festival of the wheat harvest. It seems to spring from a society in which celebrating the harvest of wheat is especially and particularly important.

These speakers of Old English also relegated to relative unimportance the widespread lord and lady terms of the cognate languages: German Herr (OHG herro) has the Old English poetic counterpart hearra; German Frau (OHG vrouwe) has the rare Old English poetic counterpart freo. At some point, dryhten came to great prominence not only as a secular term but also in Christian discourse. I find it tempting to associate this term with the more militaristic societies that arose in the sixth century. Lord and lady, however, were never displaced.

2. Possible substrate influence of the putative late Latin dialect on Primitive Old English.

The loanwords in which Old English has a palatalized (or affricative) consonant are also those in which Romance languages have a palatalized consonant in the same position. No other Germanic language with palatalization has it as early as this (perhaps circa 500 or earlier). Primitive Old English palatalization must be a post-settlement development, since it affects British place names (caito- > /kɛːd/ > ched, chet, etc., Archenfield, the River Churn, etc.). In West Germanic, only Frisian also has palatalization. How palatalization arose in Frisian is unclear to me (given the dating of the insular version, I don’t quite see how the two phenomena can go back to continental Anglo-Frisian). One could assume that palatalization was brought to Frisia by speakers of Primitive Old English returning to the continent, perhaps during the fifth-century settlement lull. But I am not competent to pronounce on any aspect of Frisian.

It seems a reasonable suggestion that palatalization in primitive Old English was a substrate influence from late Latin, carried over by Latin speakers of the south-east when they shifted to speaking primitive Old English in the fifth century, and subsequently spread to dialects of Old English outside the south-east.

This might offer an alternative explanation for the restricted incidence of palatalized consonants in northern and east midland dialects of Middle English. The received theory, by which Old Scandinavian, which lacked palatal consonants, caused Old English palatal consonants to be replaced by velars, seems shaky for several reasons. 

(1) By the time of the Scandinavian settlement (late 9th century) palatalization should have been established in Old English for 200-300 years. One would have expected affrication to have happened, i.e. palatalized /k/ and /g/ to have developed by then into the full affricates /ʧ/ and /ʤ/ and /sk/ at least into /sʧ/, if not into the fricative /∫/. One can envisage slightly palatalized velars being shifted back to fully velar articulation, but not consonants with alveolo-palatal articulation, which are more likely to be identified by non-native speakers with affricates of /s/ and /z/. 

(2) Scandinavian dialects might not yet have developed the palatals that we see in later Scandinavian, but the sequences /kj/, /gj/, and (more to the point) /tj/ and /dj/ existed and could have provided close substitutes for Old English palatal consonants, enabling those who adopted Old English speech to make the transition to the authentic English affricates. 

(3) There seems to be some evidence that words that would be expected to have shown palatalization lack it, even though they also lack Scandinavian parallels that could have ‘influenced’ the process of depalatalization.

On the theory that palatalization in Old English may have arisen from the Latin substrate, an alternative hypothesis about the restriction of palatalization can be advanced.

(1) The primary reason might be the spread of the second wave of immigration from eastern and northern areas. While, as has been suggested above, their hegemony over the whole of England brought about the diffusion of i-mutation throughout Old English, the spread of palatalized consonants, a reverse movement from the established variety of Old English into that of the later comers, may have been more limited: it was not, by this date, a living sound change; it would have been simply a substitution of one sound for another, lexeme by lexeme. 

(2) In the north of Britain there was little or no Latin substrate, and the British substrate would not have favoured the development of palatalized consonants.

If the blocking of palatalization is imagined to have occurred much earlier than the age of contact with Scandinavian settlers, i.e. from the sixth century onwards, it is easier to account for it. At this stage the palatalized consonants in lexical items spreading through Old English might well still have only been at the stage /kj/ and /gj/ (or /c/ and /ɟ/). Replacement of these by unpalatalized velars is quite explicable. The replacement of /kjir(i)kje/ by /kir(i)ke/, reflected by northern Middle English kirk(e) is straightforward.

3. At what stage did Palatal Diphthongization occur?

I suspect that there are sound reasons to question Campbell’s case arguing that palatal diphthongization occurred after consonant palatalization but before i-umlaut. It hangs virtually on the single case of the word cyse (< *ciese, in which it is supposed that æ1 would otherwise have remained unchanged through i-umlaut and been diphthongized to ea rather than ie > y). All other instances of palatal diphthongization can be equally well explained as arising after i-umlaut: e.g. /katil/ > /kjætil/ > (i-umlaut) /kjetil/ > (PD) /kjietel/ cietel is as good as (if not better than) the accepted transmission /katil/ > /kjætil/ > (PD) /kjeatil/ > (i-umlaut) /kjietel/ cietel.

But the case of words like sceað (< *skaiþi-) which require a second dose of palatal diphthongization later than i-umlaut, weighs against the argument from cyse: why not position all the PD at the same, later stage? To explain *ciese we only have to suppose that West Saxon æ1 was, like its equivalent in Anglian, originally narrower than later æ2 (i.e. /ɛː/ as against /æː/) so that following a palatal consonant it diphthongized to /iːe/, but in all other contexts, subsequently, it was lowered and merged with æ2, whereas in Anglian it was raised and merged with /eː/ (or perhaps became the raised vowel with which /eː/ of other origin merged). This does not seem inherently unlikely.

If ciepe from Latin caepa is an early insular loan (i.e. another of those later than the settlement but before i-umlaut) it would also support this argument. With replacement of intervocalic /b/ by Old English /p/, we could assume that insular Latin /kaepa/ or /kεːpa/ > */kjεːba/ > primitive Old English */kjεːpæ/ (with æ1) > */kjiεːpe/ > /kjiːepe/. This assumes that the relevant variety of late Latin kept the reflex of ae as a long vowel and distinct from the reflex of /eː/—for which I have already argued when discussing cerfille; if they had fallen together we would expect Old English */kjiːpe/ *cipe, like side.

In fact, it may be arguable that primitive Old English originally had no long close */eː/; the apparent (West) Germanic /eː/ in the small number of words like hēr, mē, etc., may be explicable in other ways, such as by lengthening of an inherited short /e/. Primitive Old English æ1 may have been the only long front vowel, which would explain how its development could be so variable between West Saxon and Anglian.

4. Failure of i-umlaut.

The form of primitive Old English which had developed among the earlier groups of settlers had no i-umlaut. This must be the case on anybody’s model. If i-umlaut was introduced from the continent by a second wave of settlers, it may not have spread consistently into all dialects. There may have been dialects that escaped it; there are quite likely to have been words and names that escaped it, which survived into literary Old English (and beyond).

There are several Latin loanwords in Old English in which a trailing /i/ or jod which would have been expected to leave a trace as the i-umlaut of the root vowel has not done so, e.g. dinor, solor, ostre, orc; and also the place names Eotol, Reculf. This is not to say that the explanation is necessarily the same in all cases.

Some of the various categories of failure listed by Campbell (§204) are unlikely to be related to this phenomenon. But names of peoples, which would be resistant to change, seem likely candidates: notably Cantware (Campbell §204 (1); his explanation ‘early syncopation’ seems the kind of argument that can be used to explain anything that doesn’t fit) as against Cent which as a place name would have been much more widely used; and Seaxe (Campbell §204 (5)), which seems a significant exponent of this failure as the name of the main people of the south and south-east.

In a future post, if I am spared, I hope to discuss the reasons why the Old English symbol y may have represented something other than a rounded high front vowel. Also perhaps some other important Christian terms in Old English.