4 From Pre-English to Old English

1.0 As was discussed in Chapter 2, the ‘tree’ and ‘wave’ models of linguistic relationships are extremely useful for linguists. However, it has already been emphasized that there are limitations to such models, which are essentially post-hoc formulations not really capturing the processual and dynamic character of language change.

1.1 The relationship of English to the other Germanic languages is not an exception to this rule. In a simple tree model such as Figure 4.1, for instance, English is a daughter language along with the others of the West Germanic group: German, Dutch, and Afrikaans, all of which are seen as deriving from a common ancestor, Proto-West Germanic, which itself derives from Proto-Germanic, which itself derives from the so-called ‘centum’ node of Indo-European languages. Such models are commonplace in textbooks on historical linguistics.

1.2 But, as was indicated in Chapter 2, such models need clarifying. The lines which connect the various nodes summarize periods of considerable complexity; such complexity is to be expected, given the limitations of tree models (see Chapter 2, in particular Section 5). Thus the lines which connect the various nodes represent periods of divergence between different languages, the so-called ‘pre-’ languages. It was during one of these processes of divergence, the ‘pre-Germanic’ period, that the sound change known as Grimm's Law took place. Similarly, we can refer to the ‘pre-English’ period of divergence which resulted in the appearance of what may reasonably be considered a discrete language. However, even this categorization may be too simple; as we shall see in this chapter, the earliest form of English, Old English, is itself a portmanteau term for a set of varieties which have distinct origins.

1.3 The speakers of what were to become the Germanic languages seem to have originated, possibly in the fifth and fourth centuries bc, in what has been referred to as ‘that bottleneck of the Baltic which is constituted by present-day Denmark and southern Sweden’ (Haugen 1976: 100). In the sixth century ad, the writer Jordanes, probably himself of Germanic origin though writing in Latin, referred to Scandinavia as vagina gentium, ‘a womb of peoples’, and this description—if extended to the north of Germany between the rivers Weser and Oder—seems to be an accurate one, even though it should be recognized that Jordanes was referring to events which took place perhaps a thousand years before he was born.

1.4 From this area of origin the Germanic peoples spread south and east; their spread to the west was constrained by resistance from first the Celtic peoples and subsequently the Roman empire. Antagonism between the Germanic peoples and the others they encountered was not consistent, warfare alternating with more peaceful contacts through trade and other forms of cultural exchange. And towards the end of the imperial period, the Romans took to hiring large numbers of Germanic mercenaries as auxiliary troops (foederati, laeti); many of the great generals of the late Roman period, such as Stilicho, were of Germanic origin.

1.5 The language spoken by the first identifiable Germanic peoples was Proto-Germanic, which is the presumed common ancestor of all the modern Germanic languages. Proto-Germanic, like all natural languages, cannot have been homogeneous, and it is likely that the differences between its dialects—which subsequently developed into distinct languages—were present from the outset. This proto-variety itself eventually split into three further groups, commonly referred to as East, North, and West Germanic. Most modern scholars are of the opinion that an initial split led to the emergence of two proto-languages, Proto-East Germanic on the one hand, and Proto-North-West Germanic on the other. Subsequently, it is held that two further proto-languages emerged from the latter: Proto-North Germanic and Proto-West Germanic.

1.6 This picture is schematic, and almost certainly a gross oversimplification, for reasons already given at the beginning of this chapter. The evidence indicates that at least the West and North Germanic groups, because of geographical proximity, continued to interact, resulting in a series of contact-induced developments; it has been said that the Germanic varieties ‘do not strive to part from one another, but remain in ever-changing connection with one another’ (Rösel 1962: 120, cited in and translated by Haugen 1976: 112). The term Proto-West Germanic is therefore, like Old English, a portmanteau expression, covering a range of varieties.

1.7 A similar narrative may be offered for the pre-English period. It is usually held that English emerged from the other Germanic usages in the first three centuries ad, deriving from a group of dialects on the shores of the North Sea with common characteristics distinct from the other West Germanic usages. It is usual to refer to these dialects either as ‘North Sea Germanic’ or as ‘Ingvaeonic’, the latter being derived from the Roman term for the tribes who lived along the North Sea littoral (cf. Tacitus, Germania). There is considerable controversy about what is meant by an Ingvaeonic language (see Nielsen 1989, 2002, for a useful conspectus of scholarship to date); most scholars hold that core Ingvaeonic languages are English and Frisian, with Old Saxon as another possible—if more peripheral—member of the group.

1.8 If this narrative were recouched in accordance with a tree model, Proto-Ingvaeonic would form a superordinate node with Old English and Old Frisian as two coordinates (along with others). Such a model does capture a truth about the relationship between these varieties, but it is of course insufficient, since it would seem to argue for the existence of some common ‘Proto-Old English’ from which subsequently the dialects of English itself had diverged. Such a picture is an oversimplification. In the remainder of this chapter, we will explore (amongst other matters) the argument that Old English arose from the interaction of varieties, previously existing as near-related but distinct dialects in old Germania, which came together in England. Such interactions seem to have been commonplace in the West Germanic period, and it is therefore unsurprising that the process should have continued after these peoples had arrived in the British Isles.

1.9 In order to demonstrate this process, one sound change will be explored in detail: the set of diphthongizations known collectively as Breaking. However, other sound changes are also brought under review, since the interconnectedness of phenomena is one of the themes of this book. These other changes include characteristic Old Anglian developments such as Retraction and Smoothing, and also changes known as First Fronting (witnessed in both Old English and Old Frisian), i-Mutation (or ‘i-Umlaut’), Back Mutation (or ‘Back Umlaut’) and Late West Saxon Smoothing (see Appendix 1). The chapter will conclude with an exploration of the epistemological issues raised by this exploration. (An earlier version of part of this chapter appeared as Smith 2002; see also Krygier 2004.)

2.0 Breaking (or ‘Fracture’) is the term generally used by Anglicists to describe a process of diphthongisation whereby, between a front vowel and certain single consonants or consonant clusters, a back glide vowel developed, at first as [u], but subsequently lowered and centred to [ə]. This back vowel combined with the original front vowel to form a diphthong. The process seems to have taken place most fully in the West Saxon variety, but it is manifested to a lesser extent in other varieties as well. Breaking is fully described and illustrated in standard grammars and narratives, for instance those by Alistair Campbell (1959: 54–60), Richard Hogg (1992: 84–5), and Roger Lass (1994: 48–51). The development is usually dated to the period immediately after the arrival of the Anglo-Saxons in Britain, a dating to which we shall return.

2.1 The term Breaking is used differently with reference to cognate languages. The term is used to describe diphthongizations in Old Frisian and Old Norse, but these are distributed quite differently from Old English within the Frisian and Norse lexicons. Thus in Old Frisian, although there are a few similarities (e.g. Old Frisian fiuchta ‘to fight’, cf. Old English feohtan), Breaking is exemplified by riucht ‘law’, ‘right’ (cf. Old English riht), siunga ‘to sing’ (cf. Old English singan), and thiukke ‘thick’ (Old English þicce). The term Breaking is applied in Old Norse studies to two changes which are treated distinctly by Anglicists, namely, Breaking and Back Mutation. Old Norse Breaking may be exemplified by such forms as Old Icelandic bjorn ‘bear’ (cf. Old English beorn ‘warrior’, with Breaking), but also jotunn ‘giant’ (cf. Old English *eoton, eoten with Back Mutation). Old English Breaking, indeed, seems to be a distinct phenomenon, as witnessed by the divergence between cognates; compare Old English weorþan ‘become’, weaxan ‘grow’ with Old Icelandic verða, vaxa, Old High German werden, wahsan, Old Frisian wertha, waxa.

2.2 The process of diphthongization seems to be triggered in the environment of following back (i.e. velar, uvular) consonants and consonant clusters:

2.3 There is also a third environment, with following /w/ (/w/-Breaking). However, the evidence for /w/-Breaking is, it is held here, problematic, and it will thus be set aside in the body of this chapter.¹¹

2.4 If, then, /w/-Breaking is left aside as part of a distinct process, the three principal environments for Breaking in West Saxon are /l/, /r/, /x/. The outputs from Breaking in recorded West Saxon, unless affected by subsequent sound changes, are forms which are spelled <ea>, <eo>, e.g. eald ‘old’, heard ‘hard’, eahta ‘eight’, seolh ‘seal’, eorl ‘noble’, feoh ‘property’. If all three environments are conceived of, quite plausibly, as in some sense ‘back’ environments, then a process of diphthongization seems phonetically reasonable (Lass 1994: 49), though there remains controversy about what the spellings <ea, eo> represent. It is usual to refer to these forms as ‘short diphthongs’, although the formulation is problematic; we will be returning to these quantitative issues in Chapter 5. For the time being the <ea, eo> forms produced by Breaking will simply be referred to as diphthongs.¹²

2.5 Breaking would seem to be a ‘realizational or phonotactic development’, as defined by John Wells (1982: 72–80, see Chapter 1, 2.4), but the question might be raised at this point as to whether Breaking is a sound change in the way it has been defined earlier— a sound change has taken place when a variant form, mechanically produced, is imitated by a second person and that process of imitation causes the system of the imitating individual to change.

2.6 The answer is supplied, in part at least, by the appearance of minimal pairs or near-pairs, albeit restricted to /rC/ environments, which arose from later developments, namely, metathesis (Hogg 1992: 303, para. 7.94) and syncopation. Metathesis yields forms in West Saxon which retain a monophthong in a Breaking environment, e.g. ærn ‘house’ (< *rænn), cf. earn ‘eagle’, while syncopation of *berern ‘barn’ produced the form bern, which contrasts with beorn ‘warrior’. Pairs such as stæl ‘place’, steal ‘stall’ would seem also to demonstrate that <æ>, <ea> represented distinct phonemes. Breaking would therefore seem to be an example of ‘split plus merger’, whereby a sound derived through a realizational development merges with a sound derived in some other way; the fact that the diphthongs produced by Breaking share in subsequent developments would support this interpretation.

2.7 But one important question, never really addressed in the handbooks, remains: why did this set of sound changes happen when and where it did, and not before or after? (See further Smith 1996: 89.) In order to address this question, we need first, following the pattern or argument developed at the end of Chapter 3, to determine the date of Breaking.

3.0 The position of Breaking relative to other sound changes can be established by typological means. To exemplify these procedures, we might investigate the relationship between Breaking and two other developments in the Pre-Old English period: First Fronting and i-Mutation.

3.1 First fronting was discussed briefly in Chapter 2, 2.10. It is traditionally described as a development whereby a → æ, except /_ C [+ nasal], [w]. The Proto-Germanic short open back vowel a appears as the short open front vowel æ in West Saxon, except in the environment of a following nasal consonant or [w] (despite Campbell 1959: 55; see Hogg 1992); thus forms such as dæg, glæd, with an open front vowel, appear beside land, with an open back vowel. i-Mutation is perhaps the most morphophonologically important of the prehistoric Old English sound changes, and its processes can be paralleled in many of the Germanic languages. The rules are as follows: V[+ back] → V[+ front]/_ $ i, j; V [+ front, + open] → V [+ front, + close] /_ $ i, j. In other words, when /i/ or /j/ stood in the following syllable, all stressed back vowels were fronted. Examples of i-Mutation include menn ‘men’ (from Proto-Germanic *manniz), yielding the Present-Day English distinction between man (singular) and men (plural). (For a fuller set of examples, see Appendix 1.)

3.2 The dating of Breaking in relation to these developments is based on an argument along the following lines; the forms eald ‘old’ and ieldra ‘older’ are relevant. The form eald seems to derive through Breaking from an earlier *æld, with *æ as the product of First Fronting, cf. Modern High German alt. The form ieldra derives from an earlier *ealdir-, which provides an i-Mutation environment; ie is the umlaut of ea, not æ (which would have been raised to e, as happened to the word in Anglian dialects, yielding Present-Day English elder). Thus the chronological sequence First Fronting–Breaking–i-Mutation is established, a typology which remains generally accepted (though see Campbell 1959: 107–8 and references there cited). Breaking was not an ongoing process; by the late West Saxon period, it no longer took place in new, borrowed vocabulary—cf. pæll ‘pallium’ (cf. Jones 1989: 54; for Jones's other examples, though, see Campbell 1959: 162n).

3.3 As is suggested by its alternative specification (‘Anglo-Frisian Brightening’), First Fronting is a change which is usually dated to the Ingvaeonic period of pre-English, before the arrival of the Anglo-Saxons in Britain, but Breaking, which is restricted to varieties of Old English and is not found in the same form in Frisian, is seen as immediately postdating Adventus Saxonum ‘the arrival of the Saxons’ (as the Anglo-Saxon invasions are commonly termed: Keynes 1999: 5–6). Breaking may therefore be considered the first ‘English’, as opposed to Ingvaeonic or West Germanic sound change affecting stressed vowels.

3.4 The circumstances of the Adventus Saxonum remain obscure; the phrase, which was coined by the early British historian Gildas and subsequently adopted by Bede, was used to describe a sudden, particular event in English history, but can be usefully repurposed to refer to the gradual process of settlement by Germanic peoples which characterized the late and post-Roman period in Britain. Bede's famous account of the geographical distribution of Angles in the north and east, Saxons in the south and Jutes in Kent and along the English south coast has been broadly confirmed by modern archaeological evidence, though there is evidence of less significant settlements by Frisians, Franks, and even Norwegians (Yorke 1999: 415–16 and references there cited).

3.5 The distribution of the various Germanic peoples on the eve of the Adventus Saxonum is both uncertain and controversial. However, most modern scholars agree that the Angles occupied the area in modern Denmark known as Angeln, to the north of modern Schleswig-Holstein, whereas the Saxons seem to have lived rather more to the south, in the area bordering Jade Bay, centred on the river Elbe and between the rivers Weser and Ems. This distribution is confirmed inter alia by the distribution of brooch types in grave goods: saucer-brooches characteristic of the Saxons appear in Southern England and in the Elbe–Jade region, whereas cruciform brooches characteristic of the Angles appear in burials in Angeln and in the English Midlands and North (see further Wilson 1981: 37ff.).

3.6 Given the principles of dialect geography, it is not therefore surprising that Anglian, though in many ways a West Germanic variety, differed from Saxon by being more like North Germanic. While still in use in the Germanic homeland, Anglian was the variety geographically nearest to the North Germanic dialects, and Anglian society evidently shared a common cultural heritage distinct from that of the Saxons. It is no coincidence that Anglian cultural sites in England, such as the Sutton Hoo ship-burial, look to Scandinavian artistic models, and that the Old English epic poem Beowulf, which seems to have emerged in the Anglian culture of the East Midlands, tells a trans-Baltic story linking Denmark with the land of the Geats, usually interpreted as located in what is now southern Sweden (see Hines 1984: passim, though see also Townend 2002).

3.7 Moreover, there is some slight intralinguistic evidence in the Old Anglian texts for an Anglian–North Germanic linguistic connection which predates the Viking invasions of the ninth century, for instance the preposition til for West Saxon tō in the Moore version of Caedmon's Hymn, dating from c.737. The verb aron ‘are’ in the tenth-century Lindisfarne Gospels Gloss is also perhaps of relevance; it is distinct from the Norse form—compare Old Icelandic eru—but clearly closer to the North Germanic pattern than the West Saxon sind(on)—compare Present-Day German sind. (See Townend 2002 for a thorough survey of English–Norse linguistic relationships during the Anglo-Saxon period.)

3.8 Since historical explanation in linguistic study depends—like (arguably) all historiography—on the observation of correspondences, it would seem that Anglian–Saxon linguistic distinctions relate at least in part to the closer placing of Anglian, during the period when the Angles still lived on the continent of Europe, to North Germanic. It would therefore seem logical to investigate whether the distinct developments with regard to Breaking in Anglian and Saxon derive from their distinct Germanic ancestries and their different original geographical locations. It is to this question that we should now turn.

4.0 We might begin this section with the most obvious distinction between Anglian and Saxon: the difference between the two varieties in /l/-Breaking environments.

4.1 To illustrate the process we might trace the development of one form: West Saxon eald ‘old’, Anglian ald. Traditionally it has been held that both forms derive from Proto-West Germanic */ɑld/, the vowel of which has been retained in Present-Day German alt. It was therefore held that the common ancestor of Anglian and Saxon—along with the ancestor of Old Frisian, the other main Ingvaeonic variety—underwent First Fronting: thus West Germanic */ɑld/ became Ingvaeonic */æld/.

4.2 A frequent assumption is that Old English was originally a single variety which subsequently diverged into various accents and dialects: Anglian and Saxon. According to this view, Ingvaeonic */æld/ underwent Breaking in West Saxon to produce the historical eald, whereas in Anglian the form underwent a distinct sound change known as Retraction to produce ald. Thus the vowel in the Anglian form underwent a pendulum shift from /ɑ/ to /æ/ and then back to /ɑ/.

4.3 Recently, however, scholars have looked again at the plausibility of this pendulum shift. Richard Hogg (1992: 80–1, esp. para. 5.13), following a suggestion originally made by Bülbring (1902), has argued that, in the ancestor of recorded Old Anglian, proto-Germanic /ɑ/ failed to undergo First Fronting in the environment of a ‘covered l’ (i.e. /l/ + consonant).

4.4 If Hogg's view is accepted—and it is certainly plausible, not least in terms of economy—then this would suggest that the realization of /l/ in the ancestor of Anglian was markedly ‘back’ in quality even before the advent of First Fronting. The question then arises: is there any independent evidence for a distinct back realization of /l/ in Anglian?

4.5 As is the case when dealing with phonetic details of a language variety of such antiquity, we have to build on a mixture of small indications, including the analysis of correspondences. With regard to velar /l/, the best and most relevant evidence comes from the Present-Day Danish of east Jutland, where a velarized /l/ is still in use (see Haugen 1976: 275 and reference cited there). In more northerly Scandinavian dialects this sound eventually merged with the acoustically somewhat similar so-called ‘cacuminal’ or ‘thick’ /l/ (see Haugen 1976: 274–8); thick /l/ is usually considered to be an ‘/r/-like /l/’ ([r, l] are of course in many languages not distinct phonemes, for example in Japanese). Haugen (976: 273) argues that a velarized /l/ was the usual realization in Pre- (i.e. pre-550 ad) and Common (550–800) Scandinavian.

4.6 It therefore seems at least possible that Anglian, the variety of West Germanic closest to North Germanic, could have developed its early velarised /l/ through contact with Pre-Scandinavian while the Angles were still living in their Germanic homeland. The velarized /l/ prevented the development of First Fronting in this environment in the ancestor of Anglian: thus the retention of ald /ɑld/.

4.7 This velarised /l/ moved with the Anglians during the invasions of the fifth century ad. In England, the velarised /l/—perceptually quite a significant feature—was adopted by the West Saxons, in a situation of linguistic contact (why West Saxons should copy an Anglian form will be discussed in 7, below). But since the ancestor of West Saxon had had First Fronting, the outcome of the velarization in that variety was somewhat distinct from Anglian in terms of vocalic development.

4.8 There remains the question of Anglian forms such as ældra ‘older’, derived from */ɑldira/, with the i-Mutation of ‘retracted’ a. That /l/ has not restrained fronting due to i-Mutation suggests either that the /l/ has changed in its realization or, more probably, that the ‘front’ vowel-harmony effects of i-Mutation outweigh or compensate for the ‘back’ quality of /l/. Such mutation effects are still recorded in Present-Day English (see Wells 1982: 533–4).

5.0 The question of /r/-breaking is a little more problematic. The distinct Northumbrian development of /r/-breaking, combined with the evidence of Present-Day Northumbrian dialects, would suggest that a ‘back’ /r/ first developed in Northumbrian varieties of Anglian, subsequently spreading—again in England—to other accents.

5.1 ‘Uvular r’ [r] develops earliest in North Germanic in Danish (see Haugen 1976: 72–3), and has spread from there into southern Sweden (on the history of /r/-realizations in Germanic, see Erickson 2002). The evidence would seem to indicate that the variety of Anglian which ultimately became Northumbrian derived its uvular realization of /r/ from the period of contact between West and North Germanic varieties. A weakened, velar form of the uvular /r/, rather akin to that found in some varieties of American English (see Lass 1983), could have been subsequently adopted by other varieties of Anglo-Saxon. Such a development would account for the early, Ingvaeonic-period failure of First Fronting in the ancestor of Old Northumbrian (producing barnum, etc.) beside the later, post-Ingvaeonic and post-First Fronting developments in more southerly varieties of Old English. If this hypothesis is accepted it would of course also suggest that the precursor of Old Northumbrian was developing as a variety distinct from the rest of Anglian even before the Angles left their Germanic homeland.

5.2 However, it is interesting that metathesized or syncopated forms, such as ærn ‘house’ (<*rænn), bern ‘barn’ (<*berern), which produced minimal pairs (see 2.6), were not subjected to breaking-type diphthongisations. This fact would suggest either that these metatheses took place after Breaking had been completed or that the conditions which triggered Breaking had ceased to be operative. One possibility, which would depend on there being a chronological gap between Breaking and metathesis, is that the realization of /r/ had lost its back quality by the time the metathesized forms developed.

5.3 Another, perhaps more likely, possibility is that the /r/ involved in metathesis was realized in a way distinct from that involved in Breaking, perhaps syllabically (as in some varieties of present-day Scandinavian, e.g. Dano-Norwegian; see Haugen 1976: 74–5), or with a glide vowel between /r/ and /n/. Such developments as the latter are frequently found as intermediate stages in some kinds of metathesis (see Samuels 1972: 16–17), and would seem a logical development of the syncopated forms as well; cf. Common Slavic *zolto ‘gold’ > Russian zoloto (East Slavic) beside Czech zlato (West Slavic). As has been pointed out, the precise mechanisms involved in metathesis have received surprisingly little attention (see Jones 1989: 191, but see more recently Blevins and Garrett 1998 and references there cited for a cross-linguistic discussion).¹³

6.0 The third, and most productive, Breaking environment is the fricative /x/. All varieties of Old English demonstrate breaking before /x/, and this would suggest that it was invariably realized in all environments, at an early date, as a ‘back’ velar fricative consonant, as in present-day Yiddish or Afrikaans and in some varieties of Dutch (see Lass 1994: 75). It would thus have differed from the modern front–back distribution seen in (e.g.) Present-Day German hoch, Hoechst, with [x, c], respectively. Such a distinction seems to have emerged towards the end of the Old English period, and it is sustained by those varieties which have kept /x/ as part of their phonemic inventory, for example the Scots distinction between [nɔxt, nict] ‘not’, ‘night’. However, the contrast between Old English feoh ‘property’ (with Breaking) and its cognate Old High German feho, or between Old English (West Saxon) seah ‘saw’ and its cognate Old High German sah, would seem to indicate that the general realization of /x/ <h> as a back consonant whatever the environment was an innovation in prehistoric Old English.

6.1 Given the argument put forward so far, can we argue that Anglian /x/ came to be realized solely as a velar (as opposed to a palatal) fricative, after the operation of First Fronting, and that it in turn affected the realization of West Saxon /x/? The evidence for such an argument is again problematic, but there does seem to be a correspondence in dating between the presumed establishment of velar realizations of /x/ in the ancestors of Anglian and Saxon and a redistribution of fricatives in what became Common Scandinavian (see Haugen 1976: 155). Could the Anglian change have been triggered by this change—and subsequently spread to West Saxon when the two varieties came into contact after the Adventus Saxonum? Certainly there would seem to be room for further investigation in this area.

6.2 However it arose, the diphthongizations were not sustained later in the Old English period. Just before the time of historical records, Anglian dialects underwent the development known as Smoothing, whereby the diphthongs produced by /x/-Breaking were monophthongized; thus Anglian had undergone the change *nēh ‘near’ > *nēoh > nēh. Now, such pendulum shifts have already been considered unlikely (see section 4, above), but Richard Hogg has offered a fairly convincing explanation on prosodic grounds (see Hogg 1992: 143–4). The result would be the rightward transfer of ‘[j]-prosody’ from the first (front) element of the diphthong to the following consonant, as a result of the obscuration of the second element of the diphthong, probably in [ə]. The fact that Smoothing fails in [x]-Breaking environments when a back vowel remains in the following syllable indicates that something along these lines had taken place, cf. the alternation þuerh ‘crooked’ (with /rx/-Breaking) beside þweoran accusative singular (Hogg 1992: 144).

6.3 The preconditions for Smoothing would seem to be two: obscuration of the second element of the diphthong, and subsequent fronting of the phoneme /x/. Something similar occurs in Late West Saxon (‘Late West Saxon Smoothing’, Campbell 1959: 131); this development has generally been seen as distinct from Anglian Smoothing, but it may simply be a later development of the same kind. There is some evidence that more southerly dialects of Old English and Middle English were more conservative with regard to diphthongal developments, and it may be that the obscuration of the second element of the diphthong, one of the two preconditions, took longer to develop here, and that Smoothing was therefore also somewhat delayed in consequence. (See also Hogg 1992: 101–6, esp. 1992: 103, para. 5.44, for the lowering of the second element of the diphthong produced by Breaking, which seems to be detectable earliest in Anglian.)

6.4 The outcome of the developments just discussed was that the Middle English distinction emerged between ‘front vowel + front /x/’ and ‘back vowel + back /x/’ in knight, nought, for example, a distinction which would only be (partially) obscured by later developments.

7.0 It will be fairly obvious from the preceding discussion that much remains obscure about the origins of the Old English sound changes, including Breaking. However, it is argued here that a reasonable hypothesis as to the origins of this sound change may be put forward, while admitting freely that final proof for such a course of events will almost certainly always be lacking.

7.1 The hypothesis depends on two insights:

7.2 The first of these points has been covered explicitly in the preceding argument. The second point has been made somewhat implicitly hitherto, and thus needs a little expansion. The history of Old English is often taken as the history of the emergence of ‘standard’ Old English, West Saxon. Yet West Saxon as we have it, it has been argued here, is really the product of an earlier interaction with Anglian, subsequent to or perhaps during the Adventus Saxonum. There is some evidence that, during the earlier phases of the Anglo-Saxon period, Anglian culture was dominant, and thus, we might expect, the Anglian variety was sociolinguistically dominant; England, after all, derives its name from the Angles and not the Saxons, and this choice seems to have been made early on. As Myres (1986: 107–8) puts it:

It has recently been suggested…that when the main tide of migration to Britain took place in the fifth century, the Angles on the Continent were already becoming the dominant element in the Mischgruppe of peoples pressing south-westward into Frisia from all the lands around the lower valleys of the Elbe and the Weser. There is no doubt that this southward pressure of Angles, Jutes, and related tribes was a major force behind the migration to Britain at this time. If in fact the Angles played a leading part in the movement, that might well account for the substitution of their name for that of the Saxons over so much of eastern Britain. It would mean that in the fifth century, as distinct from what happened in the fourth or third, the main impetus was now coming from what German scholars have termed the Grossstamm der Angeln. It would have incorporated all the restless peoples on the north German and Frisian coasts under the leadership of Angle or Jutish chieftains pressing down from Jutland, Schleswig, and the Baltic lands beyond the lower Elbe.

It is possible that the process of Anglian dominance, therefore, was already under way as the Angles and Saxons were en route to their areas of settlement.

7.3 In short, the origin of the prehistoric Old English sound changes, often treated as a rather esoteric set of formalisms, becomes explicable when seen in the context of the historically attested movement of peoples—and thus may be explained by reference to present-day observations informed by sociolinguistic theory. These sound changes may, in short, be seen as the outcome of ‘H&H’ processes of the kind discussed in earlier chapters.¹⁴

8.0 It was argued in Chapter 1 that language change, including sound change, depends crucially on contact. It was suggested there that the raw material for sound change always exists, in the continually created variation of natural speech, but sound change only happens when a particular variable is selected in place of another as part of systemic regulation. Such processes of selection take place when distinct systems interact with each other through linguistic contact, typically through social upheavals such as invasion, revolution or immigration; as Angus McIntosh has insisted, ‘Fundamentally, what we mean by “languages in contact” is “users of languages in contact” and to insist upon this is much more than a terminological quibble and has far from trivial consequences’ (1994: 137). And interestingly the set of changes known as Breaking corresponds very closely to a key moment of contact: the coming together of Anglian and Saxon varieties, during the Adventus Saxonum, to produce Anglo-Saxon.

8.1 Scholars are paid to be sceptical, and an easy response to the hypothesis put forward here might be something as follows: Where is the proof? We know so little about this shadowy period that it is very risky to attempt any explanation at all of such events as Breaking, especially based on such a problematic body of evidence.

8.2 Such arguments can be countered along the following lines. Given that, for reasons discussed in Chapter 1, historical explanation is a valid activity, what alternative is offered? Historical explanation is an argument, and absolute proof would seem to be an impossibility: thus the practice of history is a matter of ‘unfolding conversations’ between different practitioners (see Curzan and Emmons 2004).

8.3 We will return to these arguments in the last chapter of this book. In the next chapter, we will investigate, from a similar perspective, developments which took place during a considerably less shadowy period, the transition between Anglo-Saxon and Norman England.