¹ This point of spoken Latin in Roman Britain, and Jackson's interpretation of the evidence, seems to have been repeatedly misunderstood or wrongly emphasized. Of course, Jackson never claimed that the people of Britain spoke ‘Classical’ or ‘written’ or ‘literary’ Latin; or that the life of this Latin was co-terminous with the Roman occupation; nor (cf. Frere, Britannia, p. 313) that Latin was excluded from the household language of the more sophisticated classes. Jackson's views, it seems to me, are quite clearly set forth in LHEB esp. 108 ff., where he discusses (a) conservatism in the adoption of spoken forms (as also in Gaul and Spain), and the effect of the schoolmaster; (b) social class and use, and the particular role of landowners and rich merchants. For Jackson (as also for me) the written language was Latin, just as we so often see in such societies where a large world language plays the role of link with the outside. British was surely almost entirely a spoken medium, and the fact that native British names got written in the best means at hand in Latin contexts of course tells us nothing that would indicate that spoken British sentences ever got written down. The conservatism or pedantry of British Latin, among acculturated Britons, is a matter of tradition and schooling, not of reading and writing, let alone literary activities. Jackson has reverted to these matters briefly in his essay on ‘The British languages and their Evolution’ in Literature and Western Civilisation (edd. Daiches, D. and Thorlby, A.), ii The Mediaeval World (London 1973), 113 f.Google Scholar

It seems to me important to grasp the fact that such normative, conservative, even archaizing spoken use on a class basis, as applied to the imported Latin, would be a natural outcome in a society so strongly rooted in and linked with a professional institutionalized bardic (and earlier druidic) culture of learning and instruction that we know Celtic society to have been. We simply see new Latin wine in old Celtic bottles.

² This articulation of x before t and s (as opposed to the specific absorption of the nasal here) is not limited to Britain; in fact it may have been nearly pan-Celtic and characteristic of Celtic speech. It must certainly be assumed as an inter-stage for ‘Western’ Romance, i.e. that Romance which developed on Celtic ground. Phonetically we have an assimilation to the feature [-grave] i.e. to the front-tongue mid-palatal position.

³ cf. K. Jackson, LHEB (=Language and History in Early Britain), 335.

⁴ In Dacia we might have expected *samtu-. Certainly we have Aromun. (a)stimtă <stinctu-, ṭimả <icinctu-, umtu <unctu-. But the result after a seems different: Daco-Romanian sînt and even Aromun plîntă <planctu-. However, this may be later and have to do with the development a>i. In any event we can contrast defuntus and umtu ‘butter’. Also in the Balkans Albanian shows gjmtyrë ‘joint’ <,iūnctūra, but shenjt(ë) <sεnt-.

⁵ See also LHEB 93 for the further Vulgar Latin reduction.

⁶ See also LHEB 366–7, and 363, §42.

⁷ See LHEB 278 for more detail.

⁸ On the date see LHEB 576–8, §154.

⁹ If true this would be evidence requiring slight modification of Jackson's statement LHEB 402, fn. 1.

¹⁰ Precisely this is seen in RIB 1606, posuuit, listed under ‘-v- inserted’, which is normal British treatment for an inherited sequence posuit; cf. LHEB 365, §44, and esp. fn. 3. While such hiatusfilling is found in Vulgar Latin (LHEB 87) it seems more common in Britain than on the Continent.

¹¹ Pace LHEB 365, fn. 2. A situation which would fall in with this phenomenon is Welsh dryw ‘druid’ <driui(d)s (*druṷid-, etc. (see LHEB 372).

¹² Note the conservative form cited by Jackson LHEB 430; these vocalisms posited above by me are however arrived at by reconstruction from the cited forms of Mann. I note this simply for the record on a point of reconstructive principle.

¹³ On this I take a more positive view than Jackson's LHEB 645.

¹⁴ See LHEB 430–1, esp. 430 fn. 3.

¹⁵ See LHEB 307–8, §19; for Nudente see LHEB 313 on au ou eu >μ(in late 1st cent.) > ū (by end of 3rd cent.) <ṻ

¹⁶ See LHEB 272–3, §4.

⁷ See LHEB 402.

¹⁸ See LHEB 536, §126.

¹⁹ The g for c in sangto may be filling a similar office, to indicate British palatality. On the other hand, perhaps ng here represents [h], with loss of the stop as in defuntus.

²⁰ For details see LHEB 406–7, §59 and 463, §85. Jackson further points out to me that the inscription containing sinum dates from the beginning of the sixth century, and that parts of it, including this word, are of doubtful reading.

²¹ See LHEB 402–3.

²² See LHEB 365.

²³ See LHEB 364–5.

²⁴ But adquae would reflect rather the neutralization of voicing to produce homogeneous voicing in clusters, a characteristic both of Latin and of British.

²⁵ See LHEB 543 ff. and esp. 550, on this complicated question. The enormous problem in all of it is of course charting the chronology and rationale of the phonetic happenings before they became distinctive in surface (or ‘taxonomic’) phonemes. Since technical phonological theory has undergone radical refinement or even change in most quarters since LHEB was written, much of Jackson's discussion (and especially that of Martinet summarized at LHEB 710) could well be recast today in rather different frames of reference. But much of his interpretative argument remains unaffected, precisely because it is free from the changing currents of theoretical fashion. The terminology of all this is naturally relatively unimportant; thus Jackson's use of lenition for the distinctive surface result (see p. 710) as opposed to Martinet's reference to the audible phonetic beginnings is not at issue. The real problem is: granted that documents show clearly the end-result (Jackson's lenition), how early and by what criteria can we claim to see the automatic contextual rule-governed phonetics that gave rise to all this? And is there any organic relation between those phonetics and seemingly similar happenings in Latin and Western Europe? Jackson has very properly removed from any such organic relation (see LHEB 548, §132; 515, §114 ff.) the development of s >h, even though this comes later to participate in a part of the same grammatical rules.

One statement which requires modification in Jackson's account (particularly since it involves the strongest criterion we have for seeing subphonemic changes in language, i.e. intersection with a different phonological system) is that (p. 549, §133) where, adducing the testimony of early Irish loans, Jackson claims that Irish of this time would have adopted lenited British [d] [g] as Irish d g (<nt nc) and not as Irish t c (>th ch) if British lenited [d] [g] really existed then as voiced stops. He claims this on the theory that Irish d g (<nt nc) were actually closest as substitutes to the British lenited stops. But this is not at all phonetically necessary nor even likely. We know (by their non-lenition in later Irish) that these Irish sounds d g (better written dd gg) were tense and probably long. On the other hand, by the very fact of their lenition we may suppose that British lenited [d] [g] were lax and probably short; cf. the Breton analogy adduced by Jackson himself, pp. 545–6. Therefore in actual phonetic fact the closest Irish sounds to the lax British lenited [d] [g] may have been the lax (lated lenited) t c.