1 Overview

David Denison and Richard Hogg

1.1 Introduction

David Crystal estimates that about 400 million people have English as their first language, and that in total as many as 1500 million may be to a greater or lesser extent fluent speakers of English (see Chapter 9, Table 9.1). The two largest countries (in terms of population) where English is the inherited national language are Britain and the USA. But it is also the majority language of Australia and New Zealand, and a national language in both Canada and South Africa. Furthermore, in other countries it is a second language, in others an official language or the language of business.

   If, more parochially, we restrict ourselves to Britain and the USA, the fact that it is the inherited national language of both does not allow us to conclude that English shows a straightforward evolution from its ultimate origins. Yet originally English was imported into Britain, as also happened later in North America. And in both cases the existing languages, whether Celtic, as in Britain, or Amerindian languages, as in North America, were quickly swamped by English. But in both Britain and the USA, English was much altered by waves of immigration. Chapter 8 will demonstrate how that occurred in the USA.

   In Britain, of course, the Germanic-speaking Anglo-Saxons brought their language with them as immigrants. The eighth and ninth centuries saw Scandinavian settlements and then the Norman Conquest saw significant numbers of French-speaking settlers. Both these invasions had a major impact on the language, which we shall discuss later in this chapter. However, they should not obscure the constant influence of other languages on English, whether through colonisation or through later immigration. Some idea of the polyglot nature of the language (as opposed to its speakers) can be gleaned from the figures presented in Table 1.1, based upon etymologies in the Oxford English Dictionary. (Note that the already-existing language English did not get its basic vocabulary and structure from any of the languages in Table 1.1; the origins of English will be introduced shortly.)

   The OED is probably the most complete historical dictionary of any language. The languages in Table 1.1 have been chosen (from over 350 in OED!) only in order to demonstrate the variety of linguistic sources for English. The figures in Table 1.1 remain imprecise, despite elaborate electronic searches of the entire OED (with its 20+ ways of marking a French loan and 50+ for Scandinavian): exact figures are beside the point and in fact unattainable.

Table 1.1 Some sources of English words (OED²)

   Even when we are dealing with only one country, say Britain or the USA, there are a wide range of varieties of English available. These varieties are dependent on various factors. Each speaker is different from every other speaker, and often in non-trivial ways. Thus speaker A may vary from speaker B in geographical dialect. And the context of speech varies according to register, or the social context in which the speaker is operating at the time. Register includes, for example, occupational varieties, and it interacts with such features as the contrast between written and spoken language (medium) or that between formal and colloquial language.

   It will be clear that the above points raise the question of what this volume purports to be a history of. There are, we can now see, many different Englishes. And these Englishes can interact in an intricate fashion. To take a single example, how might we order the relationships between written colloquial English and spoken formal English? Not, surely, on a single scale. And as English becomes more and more of a global language, the concept of dialect becomes more and more opaque. In writing this volume, therefore, we have had to make some fundamental decisions about what English is, and what history we might be attempting to construct.

   In making these decisions we have had to bear two different aims in mind. One is to be able to give some plausible account of where English is situated today. Therefore many of the chapters pay particular attention to the present-day language, the chapter on English worldwide almost exclusively so. But this is a history, and therefore our other aim is to demonstrate how English has developed over the centuries. And not merely for its own sake, but because of our joint belief that it is only through understanding its history that we can hope adequately to understand the present.

   At this point we first introduce some conventional labels for periods in the recorded history of English. From its introduction on the island of Britain to the end of the eleventh century, the language is nowadays known as Old English (OE). From c.1100 to around the end of the fifteenth century is called the Middle English (ME) period, and from c.1500 to the present day is called Modern English (ModE). ModE is distinct therefore from present-day English (PDE), which, if a period at all, extends at most to the childhoods of people now living, say from the early twentieth century to the present. Division into periods is to a large extent arbitrary, if convenient for reference and sanctioned by scholarly tradition. There is both linguistic and non-linguistic justification for identifying (roughly) those periods, though sometimes with slightly differing transition dates, and sometimes with the main periods of OE, ME and ModE divided into early and late sub-periods. Other periodisations have been proposed, however, and in any case the transition dates suggested above should not be taken too seriously. There is no point in further discussion until more evidence of the detailed history has been presented.

1.2 The roots of English

What is English? Who are the people who have spoken it? Before we begin our exploration of the internal history of English, it is questions such as these which must be answered. If we trace history back, then, wherever English is spoken today, whether it be in Bluff, New Zealand, or Nome, Alaska, in every case its ultimate origins lie in Anglo-Saxon England. If we consider the map of Anglo-Saxon England (Figure 1.1), based on the place-names in Bede’s Historia Ecclesiastica of the early eighth century, we get some impression of what the Anglo-Saxons might have thought of as their heartland. This map is, of course, incomplete in that it relies on only a single, albeit contemporary, source. Furthermore, Bede lived his whole life at Jarrow in County Durham, and his material is necessarily centred on Northumbria and ecclesiastical life. Nevertheless, it is a useful reminder that the original English settlements of Britain concentrated on the east and south coasts of the country.

   Of course, this is not unexpected. The Anglo-Saxon speakers of English had started to come to Britain early in the fifth century from the lands across the North Sea – roughly speaking, the largely coastal areas between present-day Denmark and the Netherlands and the immediate hinterland. Bede himself states that the Anglo-Saxon invaders came from three tribes, the Angles, the Saxons and the Jutes. He equates the Angles with Anglian, the Saxons with Saxon, and the Jutes with Kentish. Certainly, it is safe to conclude that the earliest settlements were in East Anglia and the southeast, with a steady spread along the Thames valley, into the midlands, and northwards through Yorkshire and into southern Scotland.

   Looking further afield, both in geography and time, English was a dialect of the Germanic branch of Indo-European. What does this mean? Indo-European refers to a group of languages, some with present-day forms, such as English, Welsh, French, Russian, Greek and Hindi, others now ‘dead’, such as Latin, Cornish (though revived by enthusiasts), Tocharian and Sanskrit, which are all believed to have a common single source. We do not have texts of Germanic, which is usually held to have existed in a generally common core between about 500 BC and about

Figure 1.1   Anglo-Saxon England (from Hill, 1981)

AD 200. Still less is there any textual evidence for the language we call Indo-European. The most usual view is that Indo-European originated in the southern steppes of Russia, although an alternative view holds that it spread from Anatolia in modern-day Turkey. The variety of opinions can be found in works such as Lehmann (1993), Gimbutas (1982), Renfrew (1987), and the excellent discussion in Mallory (1989). Many older works are equally important, and Meillet (1937) remains indispensable.

   Whatever the actual shape of Indo-European (much work has been done to define this over the last two centuries), and wherever and whenever it may have been spoken, it will be obvious that any language which is the source of present-day languages as diverse as Hindi, Russian, Latin and English has everywhere undergone substantial change. The normal method of displaying the later developments of Indo-European is by a family tree such as that shown in Figure 1.2. Although family trees such as this are the staple diet of most books on historical linguistics, they should always be treated with caution. Indo-European is necessarily a vague, or at least fuzzy, entity, and the same is true of its branches.

Figure 1.2   The Indo-European languages

Figure 1.3   The Germanic languages

   In order to see that, consider a fairly standard family tree of Germanic, of which English is one part, such as that shown in Figure 1.3. Such a tree obscures a variety of problems, and one reason for this is that it forces a strict separation between languages which certainly could only have emerged over a period of time and where various features may be shared by apparently discrete languages.

Figure 1.4   Wave representation of Germanic (after Trask, 1996)

   It is, therefore, worth comparing the family tree in Figure 1.3 with an alternative arrangement derived from the wave theory of language relationship, where languages are placed on an abstract map according to their degree of similarity. Figure 1.4 is one such diagram, based on significant shared linguistic features – the lines marking off the spread of features are called isoglosses. What both this wave diagram and the family tree demonstrate in their different ways is that the closest language to English in purely linguistic terms is Frisian, still spoken by about 400,000 Frisian–Dutch bilinguals in the Dutch province of Friesland and a few thousand speakers in Germany, most of them in Schleswig-Holstein.

   How can we tell that the origins of English are as we have described? After all, the oldest English texts, apart from tiny fragments, date from about AD 700, and the only older Germanic texts are from Gothic, about 200–300 years earlier. And perhaps the earliest other Indo-European texts – the Anatolian languages, principally Hittite and Luwian – are from about 1400 BC. The method by which we attempt to deduce prehistoric stages of a language is called comparative reconstruction, and it is useful to consider one simple, but nevertheless important, example of this as shown in Table 1.2.

   If you compare the forms language by language, then a number of features should become clear:

•    where Sanskrit, Greek and Latin have /p/, English has /f/

•    where Sanskrit, Greek and Latin have /t/, English has /Θ/ (= OE Þ)

•    where Greek and Latin have /k/, and Sanskrit has /ś/, English has /h/

Table 1.2 An example of comparative reconstruction

   and furthermore the similarity of all the forms is so great that this cannot be the result of accident.

   If we assume that English /h/ was originally the voiceless fricative /ⅹ/, for which there is early spelling evidence, then we can note that, with one exception to the above, wherever Sanskrit, Greek and Latin have a voiceless stop, English has a voiceless fricative. The principles of comparative reconstruction then say that, all other things being equal, the earliest texts show the older state of affairs. Therefore, the four languages concerned must have shared a common origin in which the initial consonants were ∗/p, t, k/, where ∗ indicates a reconstructed form. In order to explain the apparently aberrant Sanskrit form śatám we have to claim that the original form was ∗katam and that /k/ later became /ś/. We have so far ignored the forms of who in the fourth row. Rather than explaining these here, it might be instructive to see if you can work out why the Indo-European form might have been ∗/kwis/. The example which we have just worked through, and which is called Grimm’s Law after its discoverer, the nineteenth-century linguist and folklorist Jacob Grimm, is much more complex than we have suggested. Nevertheless it may give some indication of the methods of comparative reconstruction.

   Exercises like the one just sketched form part of an edifice of scholarly knowledge built up over many years. Their success gives plausibility to hypotheses about the historical relationships between attested languages. Comparative reconstruction also allows one to fill in stages of language history for which there is no surviving historical evidence. It works most obviously in the areas of phonology, morphology and lexis, but even the syntax of Germanic and of Indo-European have been reconstructed in some detail. There is a danger that by assuming a single common ancestor one inevitably produces a single reconstructed proto-language. Potential circularity of this kind can be mitigated in ways to be discussed in a moment. In fact, much of what we think we know about the history of English is so tightly held in place in the accumulated mesh of interlocking hypotheses that its correctness is virtually certain. What appeals to the writers of this book is that there is so much still to discover.

   In this process of intellectual discovery, the linguistic data are primary, but we can anchor our mesh of assumptions by means of certain ‘reality checks’ external to the language. Some are methodological. The greater the explanatory power of a hypothesis and the fewer special cases which have to be pleaded, the more likely it is to be correct. Second, hypothesised states of the language and the necessary changes between such states are only acceptable if they can be paralleled by states and changes which have actually been attested elsewhere (the Uniformitarian Hypothesis, that the types of possible language and language change have not changed over time). Some are non-linguistic: we require our internal history of the language to fit in with what can be discovered of its external history, which in turn is enmeshed with the cultural, political, economic and archaeological histories of its speakers. (Much of this chapter is concerned with those particular kinds of relation.) And some anchors involve the histories of other languages, which have their own complex mesh of assumptions and reconstructions: when a good sideways link is found between two such language histories, each may be strengthened. Relevant examples include the values of the letters used in the Latin alphabet when it was applied to the spelling of English, and the borrowing of words at various times from other languages into English and from English into other languages. Notice that these constraints on the construction of linguistic history are as necessary for historical periods as for prehistory. Even when we have actual texts to work on, all but the most basic description is still no more than inference or hypothesis. Like all scientific endeavour, the findings of historical linguistics are provisional.

1.3 Early history: immigration and invasion

We have already noted that English is a member of the Germanic branch of Indo-European. As such it was brought to Britain by Germanic speakers. (This section has for convenience been given a rather anglocentric subtitle; after all, the Anglo-Saxon and indeed Viking invasions are emigrations from the point of view of the people(s) left behind.) Of course, when these speakers came to Britain, the island was already occupied, and by two groups. Firstly, by speakers of a number of languages belonging to the Celtic branch of Indo-European: Welsh, Scots Gaelic, Cumbric, Cornish and Manx. At the beginning of the fifth century Celtic speakers occupied all parts of Britain. Secondly, and at least until 410, there were Latin speakers, since Britain as far north as southern Scotland was a part of the Roman Empire. The withdrawal of Rome from Britain in 410 may well have been the catalyst for the Germanic settlement. In linguistic terms, obvious Celtic influence on English was minimal, except for place- and river-names (see Section ), pace the important series of articles incorporated in Preusler (1956). Latin influence was much more important, particularly for vocabulary (see Section ). However, recent work has revived the suggestion that Celtic may have had considerable effect on low-status, spoken varieties of Old English, effects which only became evident in the morphology and syntax of written English after the Old English period; see particularly Poussa (1990), Vennemann (2001) and the collections edited by Tristram (1997, 2000, 2003). Advocates of this still controversial approach variously provide some striking evidence of coincidence of forms between Celtic languages and English, a historical framework for contact, parallels from modern creole studies, and – sometimes – the suggestion that Celtic influence has been systematically downplayed because of a lingering Victorian concept of condescending English nationalism.

   As we have already mentioned, the Anglo-Saxon settlement of Britain began along the east and south coasts. The first settlements appear to have been in East Anglia. Exactly who these settlers were is hard to tell. Even the name ‘Anglo-Saxon’ is not of great help. The terms are not strictly comparable. The Angles probably formed a group of coastal dwellers in the area between, approximately, modern Amsterdam and southern Denmark (see Figure 1.5).

Figure 1.5   The homeland of the Angles

   The Saxons, on the other hand, were a group of confederate tribes which may have included the Angles. Bede also tells us of the Jutes, about whom we know little more than that. But it seems significant that Kent and the Isle of Wight, where the Jutes seem to have been based, had distinctive features of their own, both linguistic and non-linguistic, throughout the Anglo-Saxon period. Deira, in Yorkshire, and Bernicia, in Northumberland, show linguistic and other signs of having been settled by somewhat different, more northerly, groups than elsewhere.

   During the fifth century it is likely that the settlements were on the coast and along valleys, but within about a century settlement was extensive throughout the country, from Northumbria down to Dorset, excluding only the hilliest areas of the Pennines. It is remarkable how quickly the settlement of much of the country was achieved. If we are to believe Bede’s account of Hengest and Horsa, this would suggest that the first Germanic invaders came as warriors to help local British (i.e. Celtic) rulers as they fought amongst themselves. In other words, the departure of the Romans meant that the organisational structures which they had erected for the governance of the country had begun to decay. Thus a vacuum of authority and power was created by their departure, and the Germanic tribes, aware of the attractions of the country, perhaps because their fathers or forefathers had been mercenaries in the Roman army, were eager and willing to step into the breach.

   But that is not quite enough to explain the rapidity of the Germanic settlement, which was far more a conquest of Britain, linguistically speaking, than the Norman Conquest 500 years later would be. What its speed suggests is that there must have been considerable population pressure in northwestern Europe at the time, perhaps partly because in the fifth century the average temperature was lower than it had been earlier and would again be later. Whatever the case may have been, this conquest saw an overwhelmingly rapid replacement or absorption of the existing Celtic linguistic community by the newly arrived Germanic speakers. There is now some genetic evidence for mass immigration to central England (Weale et al., 2002), consistent with displacement of the male Celtic population by Anglo-Saxons but saying nothing about females. Before long Celtic speakers had been confined to the lands west of Offa’s Dyke, to Cornwall, the northwest, and north of the Borders of Scotland. The gradual elimination of Celtic has continued remorselessly, albeit slowly, ever since. It may only have been with the coming of Christianity and the establishment of churches and abbeys that Anglo-Saxon England started to achieve the beginning of the types of political and social structure which we associate with later centuries.

   After this first phase we witness the consolidation of Anglo-Saxon authority over their newly won territory in the seventh century with the emergence of what we now call the Heptarchy, or the rule of the seven kingdoms. These were the kingdoms of Wessex, Essex, Sussex, Kent, East Anglia, Mercia and Northumbria. It would be misleading, however, to think of these ‘kingdoms’ in modern terms: they were more like tribal groups, their boundaries vague and subject to change, not susceptible to the precise delineation of the kind that we are accustomed to today. Even their number, although hallowed by antiquity, may be due as much to numerology as to historical fact.

   We shall return to the issues surrounding the Heptarchy, but not the Heptarchy itself, when considering political and cultural history. At the moment we need only observe that by the later seventh century the major centres of power appear to have been amongst the northern kingdoms, and especially Northumbria. In the following century Mercia gradually became the key centre of power. But this was to change. For at the very end of the eighth century, in 793, as the Anglo-Saxon Chronicle reports, ‘the harrying of the heathen miserably destroyed God’s church in Lindisfarne by rapine and slaughter’ (Garmonsway, 1954: 56). For now Britain was to be invaded once more. This time, however, the invasions were to come from fellow-speakers of Germanic, namely Scandinavian Vikings from Denmark and Norway.