Professional Documents
Culture Documents
Wave Theory of Linguistic Change
Wave Theory of Linguistic Change
In linguistic pedigrees, proto-languages give way to language families that are in turn replaced by modern descendants, so that proto-Indo-European begat Germanic and her sisters; Germanic begat West, East, and North Germanic; and West Germanic begat English, Dutch, German, and Frisian. A similar diagram might illustrate how an original Christian unity was divided by schism into Catholic and Protestant, from which developed the related sects of contemporary Christianity. The problems with family trees in religious history are well known to Mormons: we are emphatically Christian, but we disavow any place in the offshoots of Catholicism that have appeared since 1517. What may be less well known is that family trees are just as unsatisfactory models in historical linguistics as well. The more you know about a languages history, the less family trees can describe it adequately. It seems clear enough now that there never was a cohesive speech community that spoke West Germanic, for example, but instead a fairly wide range of related dialects whose speakers regularly shifted in their allegiance to each other. A unified Germanic or Indo-European is similarly unlikely. It still makes sense to discuss and reconstruct features of proto-languages, but you have to assume considerable diversity in these prehistorical language communities. Family trees are also inadequate for describing dialects and their relations to each other. Its said that you can start walking in Portugal and continue your trip through Spain and France to the toe of Italy without ever finding a distinct linguistic border. To the extent that people speak their local dialect, they can understand speakers of the dialect across the valley without difficulty, who in turn can understand the people in the next valley over, and so on from the Atlantic to the Mediterranean. And yet someone from the middle of Spain will not understand someone from the middle of France or Italy. One attempt to deal with the inadequacies of family trees in the late nineteenth century was the wave theory. It proposed that language change emanated from centers of innovation, more or less like ripples in a pond. In some cases, one can even find documentary evidence of a linguistic feature spreading slowly up the Rhine over centuries as people adapted their speech habits to those of their neighbors. The isoglosses that define the borders of a linguistic feature extend and decline over time, sometimes joining with and reinforcing other isoglosses, even splitting one language continuum into speech communities that are not mutually intelligible. The wave theory is not perfect, and certainly not the last word in language change, but it has been a useful and stable model in dialect geography. It seems to me that something like wave theory is necessary to account for Church history. Such a theory of Church history in a nutshell might note that the religious landscape of the Restoration in nineteenth-century New England was marked by disagreement and strifebut argument is also a form of communication, and the obstacles to shifting allegiance from one Christian sect to
another seem to have been quite low. Joseph Smith appeared on the scene as a highly productive center of innovation from whom many new religious features emanated, although into an existing American Protestant environment rather than writing on a blank slate, and in competition with other voices. The isoglosses that defined Mormonism rendered it fairly impervious to influences from outside, changing it from a dialect of the second Great Awakening to a separate religion in its own right, and then the trek westward substantially isolated the new faith for nearly a century. The Church continues to be vigilant against unsanctioned innovations or relic-area conservatism that might lead to the development of schisms or differentiation into religious orders. One nice feature of the wave model is that it acknowledges the religious heritage of the first Mormon converts without requiring us to view ourselves as an offshoot of the Methodists, or something equally nonsensical. While languages may appear to be a rather static phenomena to the outside observer, those who have looked into the intricacies of a language as it was written and spoken over a period of centuries will find that not only do languages change, but they often change drastically, absorbing new words from other languages and other dialects as contact, trade, and the advance of technology bring in new ideas and new goods that all require their own unique way of being expressed. Theoretical models like the Wave Model and the Family Tree Model have been put forward for predicting and describing the changes that occur in individual languages, but even these products of research and observation have their own unique advantages and disadvantages. Despite the fact that the Wave Model was originally devised in response to the earlier Family Tree Model and the two were at one time sharply divided by strict party loyalties, most linguists since the end of the nineteenth century have considered them to be complimentary. (Smith 38) Can the two models stand alone as solid opposites? Is one more suited than the other at predicting language change, or should they in fact be used in tandem, combined into a stronger, more coherent theory that better explains both the divisions that languages undergo and the way in which new traits are diffused geographically? Proposed by Johannes Schmidt in 1872, the Wave Model is the keystone of a theory in which it is stated that language change spreads in waves outward from central points of initial change in a series of concentric circles. (Clark 328, Fox 129) The Wave Model implies that the evolution of each individual language into its modern form is the product of the spontaneous appearance of new dialectal features (within a given language or as traits of other languages) in a given area combined with the subsequent spread of these features through contact in a wave form. It incorporates the borrowing of words that happens between languages, (such as when one culture introduces new concepts or trade goods to another) the borrowing of new linguistic features by children from their peers, and the spread of those children (with their modified dialects) to other areas (Smith, 38). Taken at face value and employed exclusively, the Wave Model ignores direct lineage and instead focuses entirely on the component parts of a language. Instead of drawing a hard line that shows one language leading to another in an unerring progression that ignores dialects and borrowings from other tongues, the Wave Model describes each language in much more detail, creating a map or diagram which tries to show the interrelationship of tongues without resorting to the direct lineage tactics of the Family Tree Model. This map of change and influence would appear at first glance as a mass of intersecting circles merging together and sliding apart, but each blending illustrated in such a map indicates a change that has occurred over time between any given language and another, creating new languages and new dialects
from the blending of multiple parent tongues. These blends are, in truth, the strongest testaments to the statement that a language like English is in fact the daughter of not just a West Germanic Anglo-Frisian Dialect, but of a whole mess of unique languages and dialects within that category as well, including Central French, Norman French, Old Norse, and countless other tongues which have worked their way into any particular dialect of our current English over time, even if only in the form of a single word (like boomerang, which is from an Australian Aboriginal dialect, or karaoke, which is Japanese.) In our modern environment of rapid, global information exchange and a regular force of travel and tourism all over the world, Schmidts Wave Model still functions in part, but it begins to fall apart when faced with the displacement of dialectal traits that happens when one person visits or moves to a foreign country, or with the seemingly random way in which changes spread over the internet, emphasizing the fact that, in truth, not all change occurs in concentric circles or waves. As it stands, the Wave Model cannot explain the means by which some changes spread from multiple sources or stop short in the face of natural boundaries (rivers, lakes, mountains, etc.) or the way in which changes in dialect which occur through the medium of the internet (which features both written and spoken information) effect individuals in a completely random fashion (geographically speaking) and may or may not transcend from a written form to a spoken medium or vice versa. True, it could be said that, if we were to draw up a network map of cyberspace, (featuring linguistic continents like Myspace, Facebook, and AOL,) it might be possible to track linguistic changes to find that they do spread in a fashion similar to that of the Wave Model, but across a completely different topography with its own natural boundaries, walls between minds, walls created by a difference of language, by shyness, censorship, and intellectual bias that kill the spread of new dialectal changes before it can even occur. When comparing what we know of the historical record in regards to language change (that which we have sufficient information to track and analyze) to the expected outcomes that we can formulate using Schmidts Wave Model, we find that the idea of diffusion between languages to build new languages by a sort of blending actually holds a fair amount of water on its own. Not only is language contact a constantly-occurring process, especially in todays world of quick, electronic communication between people worldwide, but new words are continuously being introduced into other languages as these contacts lead to a transfer of concepts or products that must be named (consider the Spanish: Computadora for Computer or Blog, short for Weblog, which has found its way into a number of other languages from English.) The Satem/Centum division (a division between languages that pronounce an [s] consonant or a harder [k] sound in the same place, as in Satem and Centum, two words for hundred) while originally noted as evidence for the Family Tree Model, as it originally created a neat division of the languages of the east from the languages of the west, has ultimately become strong evidence for the Wave Model with the discovery of languages that do not conform to this neat divide. (Renfrew 107) The fact that there are a number of shared traits between IE language groups that occurred after language division occurred (such as the fact that the past tense is based on the original perfect in both Germanic and Latin or the fact that some case-endings with -m instead of -bh occur in both Germanic and Baltoslavic tongues)(Fox 132) presents a significant argument, as does the widespread overlapping of isoglosses (Fox 129) which clearly indicates that not only do linguistic traits spread independent from one another, but they can come together to effect the same area in their steady march outward, creating areas that feature both dialectal changes while areas beyond feature only one or the other of the two. We do know, Says William Labov in his
book, Sociolinguistic Patterns that the growth of the affected area [i.e. the area affected by the dialectal change] may be checked by linguistic factors (Herzog 1965) or by social factors or historical discontinuities (Bloomfield 1934:344) or by the negative prestige of the group as a whole (as in the case of New York City.(Labov 319) In 1871, one year before Schmidts Wave Model was first formulated, a German Linguist by the name of August Schleicher proposed a model for language evolution that follows a similar line as the work of Charles Darwin by creating a sort of family tree to show descent and relatedness, with branches ultimately leading back to the trunk of an extinct common ancestor whose diversification gave rise to all of the many daughter forms we know today. (Clark 326) It is an overly simplistic model composed of successive comparisons, creating proto-languages as we go, until we have grouped all the languages together, creating a hierarchical tree of relationships (Fox 122-123) and in its over simplification, it makes a number of faulty assumptions, including what is probably the most grievous that the Proto-Indo-European trunk of the tree existed in only a single, uniform dialect with no prior ancestor form. (Fox 133) Unlike the Wave Model, the Family Tree Model focuses wholly on direct, one way relationships, and tends to lump all dialectal forms of any given language into one single language category if the dialects are all more or less mutually intelligible. It implies that language change is essentially a matter of language or dialect split, so that a language may have more than one descendant, but only one parent (Fox 123) whereas the Wave Model implies a sort of diffusion or bringing together, an intermeshing of traits from numerous languages and dialects that proves, much as the flesh under any given skin color does, that there are many more similarities between us as people than there are differences. When placed side by side, it becomes clear that both the Wave Model and the Family Tree Model have their uses, their strong points, and their inevitable pitfalls. While the Family Tree focuses primarily on the divergence and downward division of languages, drawing hard, unswerving lines that chart the descent of what has occurred in the most simplistic of approaches, Schmidts Wave Model is oriented more toward tracking the diffusion of traits outward from a central source, following how they migrate, change and interrelate, creating hybrids that are recognized for their numerous and mixed parentage.(Bynon 194) It is the Model not of what has already happened (in the broadest sense), but rather of how it happened and of how it may happen again. It is a tool for predictions, showing us how the new dialectal changes of today or the as yet unimagined traits of tomorrow may spread and change our languages, whereas the only prediction you can really draw by using the Family Tree Model is that any given language may one day sprout its own daughter forms, if it does not go extinct first. We could easily prefer a combination of the two models says Colin Renfrew, where there was some splitting from an original proto-language, involving some displacement of people, followed by the addition of subsequent effects by the wave mechanism(Renfrew 106). Together, the two models can be used to offer a feasible way to explain exactly how the divisions we see in the Family Tree Model might have happened, what separated one tongue from another so much that the two became no longer mutually intelligible, indeed how the very diffusion of changes triggered the division between dialects into new daughter tongues of a parent language. There is certainly proof that languages evolve and spawn in a linear fashion very much like a Family Tree pattern, one feeding into another, and although I find myself most convinced by the
much more in-depth way of looking at things incorporated into the Wave Model, I feel that its safe to say that only when the Wave Model and the Family Tree Model are put together in a sort of hybrid do they explain and allow us to predict (to a degree) the features of language change in the most reasonable fashion. What I would like to propose is a new Model, one that charts language change in depth (as per the Wave Model) in a three dimensional form which describes not only geographical information, but also the passage of time, down to the level of dialects, so that we can see exactly who talked how and when, similar to what William Labov mentions when he suggests a new version of the wave model of linguistic change which would show symmetrical distributions through time, space, and society.(Labov, 266) Granted, a lot of the data necessary for such a detailed Model has been lost over time, but with technology where it is now, I could easily see this model going into effect with our current level of information and paired up with widespread observation and data collection to map out and retain a list of the changes that take place in our languages today, tracking how change develops and spreads in an attempt to be able to draw new and more flawless predictions about the future of each language spoken in the world today. Wave theory Modern popular presenters of biological evolution do not often make a comparison between it and the development of natural languages, although such prominent figures as Charles Lyell and Charles Darwin did so.1 A recent book by Robert T. Pennock2 presents such a comparison anew with the claim that this is quite significant-the book has been highly praised for this by some evolutionists. The basic idea, as presented by Pennock, is that all languages have descended from 'a single original language (or perhaps a few)'3 and languages have developed through gradual uniform changes: 'Even though the record of transitional sequences is quite incomplete, the evidence taken together clearly supports the thesis that the different linguistic kinds we now observe arose through gradual transformation from earlier, ancestral forms'. 4 Also, the processes involved are supposed to be naturalistic and undesigned: '[Natural] languages . . . developed into and from one another over time, through piecemeal construction and unplanned transformations. . . . Though not perfectly Darwinian, this ongoing transformation is clearly an evolutionary process and a close analogy to biological evolution in its most significant aspects'. 5 Pennock makes several references to the Indo-European language family (defined below) to support his claims. This paper presents an overview of the most important kinds of change which have happened within the Indo-European language family, and shows that the comparison of language development with biological evolution is completely invalid. No previous knowledge concerning
the development of languages is presumed, and it is hoped that the paper will remove some popular misconceptions about the history of some of the major European languages.6 Emphasis is placed on the Indo-European language family for three reasons: 1. Several of the languages of this family will be familiar to many readers; 2. This family is a good representative for the most important kinds of changes that have happened universally; and 3. Most popular presentations by evolutionists have been based on examples from this family (including by Darwin). Note that there is no attempt in this paper to give a detailed examination of the biblical account of the confusion of tongues at Babel or the dispersal of the various peoples thereafter; neither is there a study of the evolutionary theories concerning the origin of language itself.7 Rather, the paper will only concentrate on the important changes which have occurred within languages and the proposed comparison with biological evolution. It is appropriate to summarize here the main arguments of the paper concerning the kinds of change which have happened:
Natural or undesigned processes of change have only caused simplification or loss of internal structure in languages. All growth of languages or increase of their complexity is a result of conscious and intelligent input, including design in several cases. The massive crosscurrents between languages (borrowings of various kinds8) are an enormous factor in their histories, and have no analogy in classical biological evolution. Great changes in some major languages have mainly happened in several short periods, which is much faster than commonly supposed. Classical Greek pher-o pher-eis pher-ei Classical Latin fer-o fer-is fer-it fer-imus fer-itis fer-unt Old High German bir-u bir-is bir-it bir-ames bir-et bir-ant Old Slavonic ber-a ber-esi ber-etu ber-emu ber-ete ber-atu
English I bear
Sanskrit bhar-ami
thou bear-est bhar-asi he bear-eth we bear you bear they bear bhar-ati
The languages of the world have been classified genetically, and languages appearing to have a common origin have been grouped together into a family. In this paper we shall concentrate on the Indo-European9 (IE) language family which consists of a large number of languages. All are presumed to have descended from a single common ancestor language known as Proto-IndoEuropean (PIE). Languages have been included within the IE family not only because of the huge number of similarities between particular words, but also because of the regularity of correspondences in sound and morphology.10 To demonstrate this, Table 1 gives the verbal forms of the verb 'to bear' in several older languages of the IE family (with the archaic English forms used to show further similarity).11 Similarity in inflexions is stronger evidence of common origin than common vocabulary, as words are easily borrowed and assimilated, but it is rare for whole morphological systems to be borrowed from one language into another. Figure 1 gives a standard diagram of some of the major IE languages in a family-tree style. This model is useful for gaining a basic overview of the IE languages, but it is widely considered to be inadequate and deceptive in a number of ways. It ignores the multitude of dialects12 which have existed, and thus omits very many branches. Most importantly, there have been many occurrences of massive borrowing by one language from another, so in terms of vocabulary there should be, for example, a very thick line from French to English and also one from Latin to English, as we shall see below.
Figure 1. Indo-European family tree with major languages. To account for the initial divergence of the branches of the family, Johannes Schmidt proposed in 1872 his Wave Theory (Wellentheorie), according to which 'changes begin in a specific geographic area and spread out concentrically from that point like waves created when a pebble is dropped into a pool'.13 'The extent of the spread then depends on the extent of the intercourse between neighbouring areas. This theory is probably a simplification of the facts, but is unquestionably nearer the truth than the family tree theory'.14 This wave theory has no parallel in biological evolution, since living creatures do not pass on acquired modifications to other living creatures near them in such waves. So we see immediately a major problem with Pennock's proposed comparison. Note that even if the family tree diagram happened to represent an accurate picture of the development of the IE languages, it would not necessarily represent the physical descent of the speakers of these languages accurately. Many peoples have adopted another language (e.g. the Normans15). Also, a huge number of people around the world whose native language is English do not have predominantly Anglo-Saxon ancestry.
Whatever the correct details of the origin of all the IE languages may be, the enormous number of related words and morphological similarities still show that all of these languages must have some form of common descent.16
Phonetic transformations
This section examines some of the many known phonetic transformations (sound changes) which have occurred within the IE languages, and the relevant implications. These transformations are amongst the most important changes which have created significant differences between related languages. The Germanic17 languages are widely Latin English Latin English presumed to have descended from the soffbcalled Proto-Germanic language. English is a pWest Germanic language, which means that pater father fagbeech all its basic vocabulary and grammar have ferbear descended from Proto-Germanic. The famous piscus fish Grimm's Law (or the First Germanic Sound pedfoot fractbreak Shift) was first described in 1818 by the Dane plenus full frater brother Erasmus Rask, and then formalized in 1822 by the German Jacob Grimm (of fairytale gc-/kchfame). This law explains how the consonants gelidus cold canhound of the whole Germanic group consistently shifted from those of PIE. This genu knee capithead transformation occurred more than two casa thousand years ago. Table 2 gives a genkin house (hut) simplified presentation of some of the consonantal shifts which occurred, using gnosc- know cordheart Latin and English words. Latin words dttthapproximately represent the original PIE 18 consonants, while English cognates decem ten tenuthin approximately represent the consonants of torrProto-Germanic.19 It is very important to denttooth thirst (parched) understand that these English words have not descended from Latin (or vice versa), but that domus timber tres three the members of each pair of words share a (house) common source. Indeed, the fact that a duo two tu thou shifting has occurred, shows that the English words have not been directly borrowed from Table 2. Examples demonstrating Grimms Law. Latin. The dialects of the German language can be divided into two groups according to the geography of Germany: Low, in the low-lying plains of the north, and High in the higher land of the centre and south. The effects of the shift described by Grimm's Law were present in all the German dialects. But within the approximate period AD 500-700,20 most of the consonants of the High German dialects again shifted consistently, the change having arisen 'in the south in the alpine
area, presumably as the result of contact with speakers of other languages'.21 This is called the High German Sound Shift (or the Second Germanic Sound Shift). High German later spread its influence to become Modern Standard German (with Luther's Bible translation playing a significant role). The Low German dialects, as well as Dutch and English, did not undergo this shift. Table 3 demonstrates the shift by listing some modern West Germanic words. The consonants of English and Dutch clearly match, while the German consonants have shifted consistently. During the Great Vowel Shift of English, which occurred between about AD 1400 and 1600, the long vowels of English shifted to their current pronunciation in England. This is why the long vowels in l, ln and ln are now pronounced in English approximately as in lay, lean and line, respectively, while in most Continental European languages these are still pronounced approximately as in lah, lane and lean, respectively. English Dutch German English Dutch German ddo drag drink dream -tbite eat hate let -kbreak ddoen ttun tten tongue tooth ttien tong tand twee -phelpen hopen slapen aap ppijp zzehn Zunge Zahn zwei -fhelfen hoffen schlafen Affe PfPfeife The spoken form of Latin which was widespread in the Roman Empire is known as Vulgar Latin, which differed from the literary Classical Latin in that many slang words were used and the word endings were simplified. After the fall of the Roman Empire, the dialects of Vulgar Latin developed to become the separate Romance languages22 and once again, the phonetic divergence involved consistent patterns of shifting. Table 4 lists some Latin words and their French, Italian and Spanish derivatives, showing just a few of the shifts.
dromen traumen two -tbijten eten haten laten -k-ssbeissen essen hassen lassen -ch-phelp hope sleep ape ppipe
Many other laws for phonetic transformations within the IE make maken machen pepper peper Pfeffer languages have been developed, and one book indeed lists 41 phonetic speak spreken sprechen plough ploegen Pflugen laws for IE languages!23 These stroke strijken streichen pound pond Pfund transformations have led, in fact, to Table 3. Examples demonstrating the High German Shift. some of the most distinguishing characteristics of the different branches of the IE family (e.g. the 'soft' palatalized consonants in the Slavic languages). breken brechen A different kind of divergence occurs when the accentuation of words changes. The Classical Latin demonstrative pronoun ille, illa, illud ('that') developed to become the definite article ('the') in the Romance forms. The different forms have arisen from the way the Latin words were accented in the different regions. For example, in the case of the Latin ille ('that,' masc.), the
French le came from the second syllable while the Italian il and the Spanish el came from the first syllable. All these transformations cannot increase the inherent complexity of a language. If distinct sounds are merged or sounds are dropped, then simplification occurs. The reasons why such shifts have occurred are not always clear, but it is certain that they have happened often, in relatively short times (about 200 years each for the Great Vowel and High German shifts), and the number of speakers does not have to be small. This shows that it is quite plausible that earlier divergences from PIE happened easily and rather quickly. Also, one key evolutionary idea is that major change is a result of haphazard naturalistic processes, and Pennock compares natural language to a 'jerry-built jumble'.24 But phonetic shifting is a major process which is unconscious and undesigned, yet is still very regular.
Simplification of inflexions
Latin A universally observed phenomenon of all language families is that inflexional morphology10 has simplified over time. The history of the IE family overwhelmingly illustrates this. Theoretical reconstruction suggests that PIE had three genders, eight noun cases and three verbal aspects.25 The evolutionary model is at a complete loss to explain why or how the complicated case system of PIE came into being. In this section, we will examine various simplifications which have occurred, and their causes. Meaning French chchateau chevre cheveu ffils fille faire plplan plaie plein Italian ccastello capra capello ffiglio figlia fare pipiano piaga pieno Spanish ccastillo cabra cabello hhijo hija hacer llllano llaga lleno ccastellum castle capra she-goat capillus hair ffilius filia facire plplanus plaga plenus
Synthetic languages are those in which several grammatical units tend to be composed into one word, and inflexions (word endings) are used to indicate the grammatical relationships between the words. For example, Classical Greek, Classical Latin and the ancient Indian language Sanskrit are all synthetic. In contrast, analytic languages are those in which each basic grammatical unit is usually expressed by a single word, and word order is very important for the conveying of the meaning.26 Modern English is very analytic, so in the clause 'the boy loves the girl', for example, swapping the two nouns would change the meaning of the sentence completely. Over time, synthetic languages have become more analytic, with the effect that inflexional morphology has repeatedly been simplified.
'Fixed word-order began to appear within the inflected languages simply as a result of growing orderliness of thought.27 Relating particles were at the same time added to inflected words wherever the inflexional meaning was vague. After word-order had acquired functional value, and the more precise relating-words were current, related endings lost their importance, and would become assimilated, slurred, and dropped, from the natural tendency of speakers to trouble themselves over no more speech-material than is needed to convey their thought'.28 The very free word-order found in Classical Latin and Greek literature (in which related words in a sentence can be separated by a long distance!) is very artificial. It is widely believed that the word-order used in speech, even during the Classical period, was rather close to that of modern analytic languages. 'We do not know the exact nature of the word-order which Cicero used when bawling out to his slave; but there can be little doubt that it was as fixed as that of colloquial Italian'. 29 Another cause of language simplification is the tendency to drop inflexions to facilitate communication when two peoples speaking similar languages mix. Anyone who has tried to speak a foreign language (or hears a foreigner speak their own language) knows that the word endings are the most easily confused or omitted elements of the words. The earliest form of English, known as Old English or Anglo-Saxon (c. AD 450-1150), was highly inflected, with three genders and several cases. Within the approximate period AD 800-1000, there were many Scandinavian invasions into England, and for a while most of NE England was ruled by Danes and this area was known as the 'Danelaw'. The language spoken by the invaders is known as Old Norse (from which modern Danish, Swedish, etc. have descended), and was similar to Old English in many ways, being also a Germanic language. Because of the mixing of these peoples whose languages had similarities, the inflexions of Old English were worn down.30 This shows that language change can happen rather quickly under certain external social conditions, and greater simplification results from greater interaction. In contrast, isolated languages seem to simplify more slowly. For example, the Icelandic language is still very inflected (with four cases) and has changed little since the Old Norse of c. 900 AD, because of the isolation of its speakers.31 This is the opposite of what is proposed for biological evolution: '. . . from what we know of evolutionary mechanisms, speciation events are likely to occur in isolated populations, and competition will quickly eliminate the less fit of closely similar forms'.32 The many inflexions of PIE have only simplified or disappeared in its descendants. For example, counting numbers of cases, Classical Latin had six, Modern German has four and the Romance languages have none. The only noun inflexion preserved in Modern English is the possessive ending ''s' which is a survival of the common Germanic masculine singular genitive case ending.33 This was the chief way of expressing possession in Old English, so the Old English version of Matthew 12:42 (with modernized words) is 'Southland's queen . . . came from earth's ends to hear Solomon's wisdom'.34 The alternative analytic possessive construction the queen of the South, etc. is thought to have arisen in English through French influence.35 For the verse just quoted, Wycliffe already uses 'of' forms in his 14>century English translation of the Bible.
The grammatical gender of nouns has been completely eliminated from English. The Romance languages have lost the Classical Latin neuter gender, while Dutch, Danish and Swedish have merged masculine and feminine to form the common gender. In summary, natural processes have only caused languages to become more analytic over time and caused inflexional morphology to be simplified. English grammar is the greatest example of the effects of these processes. All evidence strongly suggests that there was a complex beginning for the language families of the world (not just for the IE family, in fact). Thus the evolutionary model has an enormous problem in that it postulates a gradual transition from simple to complex, yet the observed processes are always going the wrong way. The following three sections examine the main ways that languages have grown, and show that these are certainly not the product of naturalistic processes.
Word formation
Word formation is a very common way of creating new words within a language. The two main kinds of word formation are compounding, which involves joining simple words together (e.g. 'sun-light', 'up-right'), and derivation, which involves adding affixes (prefixes, suffixes or infixes) to existing words (e.g. 'e-volu-t-ion'). Compounds seem to have always been a significant feature of IE languages. Compounds like 'whitehouse' were apparently even a common feature of PIE.36 The multitude of compounds and derivatives in English is amazing. The native Germanic root bear has given over 40 derivatives in English and the Latin root ced/cess- (meaning 'to go') has given over 80 derivatives (mostly borrowed from Latin). 'We share with French our most elaborate derivative in-com-pre-hen-s-ib-il-it-y, with its root 'hen' and its eight affixes and infixes'. 37 The invention of new compounds continues in modern times. Philosophers indulge in such specimens as 'the in-ness of a one-ship which fills the us-dom with anti-ty'.38 As one writer has pointed out,39 English has become very analytic in its grammar but very synthetic in its vocabulary because of the abundance of compounds and derivatives! That is, the internal structure of the words themselves is complex, and is certainly the product of much human intelligence and design, not naturalistic processes.
Semantic modification
Another very important way in which a language grows is by semantic modification of existing words. The meanings of most words in many languages have changed greatly during their history. Narrowing of meaning has often occurred. For example, the English word 'starve' was the usual general word for 'to die' in Old English but was displaced by the Scandinavian word 'die'; thenceforth 'starve' narrowed its meaning. Similarly, the English word 'deer' was originally the general word for 'animal'.40
Abstract words form the bulk of the vocabulary of major modern languages like English. Although languages have several native words which express simple abstract concepts literally (e.g. the native Germanic English words 'love', 'live', 'feel', 'think'), the most common way of forming abstract words is by metaphorical extension of existing words. The renowned English etymologist Ernest Weekley wrote: 'Every expression which we employ, apart from those that are connected with the most rudimentary objects and actions, is a metaphor, though the original meaning has been dulled by constant use. Thus, in the above sentence, expression means what is squeezed out, to employ is to twine in like a basket-maker, to connect is to weave together, rudimentary means in the rough state, and an object is something thrown in our way'.41 Note that all of the words just listed were borrowed into English from Latin (some through French) and they already had their metaphorical senses in Latin-no change of meaning occurred in the borrowing. By the use of such abstract compounds, Greek and Latin were developed to great complexity of expression. There is hardly an abstract concept or thought which cannot be expressed clearly in these classical languages-much philosophical writing was in Greek. A similar phenomenon occurred in Old English, in which very many abstract words were formed by compounds of native Germanic words, instead of by borrowings from Latin.42 Compound abstracts involve two design aspects: the construction of the compound and the semantic extension to a metaphor. The metaphorical senses were obviously originally designed consciously (though naturally), so human intelligence has played a significant part. Slang mostly consists of colloquial metaphor. A very large number of slang words in Vulgar Latin became the proper word for a common object in the Romance languages. A popular example is the Classical Latin word testa which meant an earthenware pot.43 In Vulgar Latin testa became a slang word for 'head' and finally became the normal word for 'head' in French, now spelt tete.44 Yet again, the development of slang meanings is a result of human intelligence (often subtle or sarcastic). The later transition of a word from a slang meaning to a formal meaning may not be very conscious, but the initial slang senses of words are used quite deliberately. A final illustration of the huge number of semantic extensions which have occurred in English is the word 'stock'. The Oxford English Dictionary gives 65 different meanings of this word (with very many sub-senses)! In summary, the great growth by semantic extension of existing words is obviously a product of human intelligence alone.
Borrowings
A major theme of this paper which cannot be overemphasized is that borrowing has played a huge role in the growth of the major IE languages. In this section we will briefly examine some important kinds of borrowings which have taken place in history. Once again, these all involve conscious and intelligent input. English supplies the greatest example of large-scale borrowing. As a result of the Scandinavian invasions mentioned earlier, English borrowed a large number of Old Norse words.45 But the Scandinavian influence on English is small compared with the French influence. For almost three centuries after the Norman Conquest of England, two languages were spoken in England: French at the court and in the upper classes, and English by the common people. All the kings of England in this period spoke French as their first language! Many people were bilingual, and English consequently borrowed a huge number of French words. The next stage in English borrowing consists of the very interesting 'learned' borrowings. The Romance languages lost the bulk of the Latin abstract words when society went into the Dark Ages. The mediaeval vernacular languages (such as the various dialects of English, French and German at that time) lacked the necessary vocabulary for discussion of abstract questions (e.g. in theology), so Latin was used predominantly. However, towards the end of the Middle Ages, writers started to borrow abstract words from Latin into their respective vernacular languages (and later from Greek, particularly for scientific and medical terms). After a while the vernacular languages were enlarged greatly by such 'learned' borrowings on a massive scale, though many writers still used Latin because they had a contempt for the 'Latinization' of the vernacular languages.46 The entry of these Latin words is sometimes called learned transmission. The 'native' words in the Romance languages have descended naturally from Vulgar Latin by popular transmission, which involves the particular phonetic shifts peculiar to each Romance language.47 But 'in contrast to popular transmission, learned transmission is instantaneous, voluntary, and modifies the orthography and pronunciation of the Latin word as minimally as possible'. 48 Doublets are two words from the same source which enter a language in different forms at different times. Each of the Romance languages has a large number of doublets, because the one Latin word has descended naturally by popular transmission, and then entered again later by learned transmission. As English has borrowed many French words and also their Latin originals, English has many of these Romance doublets, some of which are shown in Table 5. One can view the vocabulary of English as consisting very roughly of several 'strata': 1. 2. 3. 4. 5. West Germanic native words, Scandinavian and then French words from the successive invasions,49 learned Latin and Greek words, and finally modern borrowings from areas all round the world.
Latin Origin
Table 5. English doublets of French/Latin origin. A very significant fact is that the Latin words in the learned stratum are in an older form than their popularly transmitted French derivatives, which are in the stratum below! Similar layering structures occur in other modern IE languages. This emphasizes that the picture of language history is very complex and not like a simple family tree with gradual divergence along separating lines. A calque50 or loan-translation is a borrowing of a compound word from another language where each component is translated into native words and then joined together. While English and the Romance languages have borrowed most Latin words in their original form more or less, some other languages abound with calques, of which German is a prime example.51 Table 6 lists some German compounds and gives the literal meanings of their components and their English translations, which themselves are borrowings from Latin. For these English words the Latin components match the German components exactly in meaning. Similar calques occur in many other languages, including many in Classical Latin, derived from Classical Greek. All these calques emphasize the large role which intelligent design has played in the history of languagescalques obviously involve even more conscious design than direct borrowings. Note that significant borrowing from Latin had already occurred long before the period of learned borrowings. For example, early Germanic tribes borrowed Latin words extensively from the Romans because of trade and similar kinds of contact. As a result, the German language has a surprisingly large number of very 'German-looking' words which happen to be Latin words borrowed early and which later underwent the High German Shift52 (and must be distinguished from later learned borrowings and calques). One author lists 89 such German words which were borrowed in this early period from the Romans!53 German an-ziehen aus-schliessen uber-leben Literal Meaning Latinate English on-draw out-close over-live at-tract ex-clude sur-vive (French)
Since the Middle Ages, there has also been large-scale borrowing into the major IE languages from many languages around the world, not just from other European languages. Borrowing also follows cultural development closely: most musical terms have been borrowed from Italian into all languages, while several shipping words have been borrowed from Dutch into English,54 as the Dutch were masters of sailing.
voran-gehen
before-go
wider-sprechen against-speak Aus-druck Aus-nahme Um-stand Zu-fall out-pressing out-taking around-standing unto-fall
In summary, extremely few words are coined anew, without being based on something previously existing.55 All the major IE languages in Europe grew enormously in their size and scope by borrowing from Latin and Greek (whether literally or through calques), and this growth was parallel to the great cultural developments during the Renaissance and the Reformation. That is, this massive growth was a product of intelligent input alone, and was certainly not a result of gradual naturalistic development within each language independently.
Loss of vocabulary
Loss of vocabulary has often occurred within languages. 'A large proportion of the rich Old English vocabulary is gone [from Modern English]. Estimates vary; most assume that between 65 percent and 85 percent of the Old English lexicon has been lost since Old English times'.56 The massive replacement of Classical Latin words by slang equivalents in Vulgar Latin, which became permanent in the Romance languages, is another example of a great loss. Loss of vocabulary may seem to be like natural selection, but is only a loss. Also, many of the causes of loss are not due to 'utilitarian fitness' (as postulated by biological evolution). Some of these causes are: abrupt displacement by the language of invaders, one synonym gaining the ascendancy over others and changes in culture or fashion.
Intermediate forms
Creationists have repeatedly criticized the theory of evolution because of the lack of intermediate fossil forms. Pennock claims that intermediate forms have been extensively inferred within language families so that, by analogy, theoretical inference of intermediate forms is reasonable within biological evolution.57 However, very large numbers of intermediate forms have actually been found for earlier forms of languages such as the Romance and Germanic languages (from German inscriptions and written records), and this has been a huge factor in the construction of the early histories of these languages.58 On the other hand, the etymologies of very many words are still disputed. Unless the history of a word's spellings and uses is well documented, or good regular phonetic laws are
shown to apply, etymology is just guesswork.59 Furthermore, there is a much greater objection to this supposed analogy. Even the greatest change of one word to another involves only a small number of letters, so postulated intermediate words hardly differ from the known words. Yet the intermediate forms which have been proposed for biological evolution bridge enormous gaps and involve millions of items of change at the biochemical level, so the comparison is unfair.
when we see what [little] amount of change 2,000 years has been able to produce in the languages of Greece and Italy or 1,000 in those of Germany, France & Spain we naturally begin to ask how long a period must have lapsed since the Chinese, the Hebrew, the Delaware . . . had a point in common with the German & Italian & each other.-Time! Time! Time!-we must not impugn the Scripture Chronology, but we must interpret it in accordance with whatever shall appear on fair enquiry to be the truth for there cannot be two truths'.61 Indeed, we must not, and need not, impugn the Scripture chronology, nor apply some liberal interpretation to it! These specified periods of years have produced vast amounts of changes in these languages! As we have also seen, very many words we use daily are not 'battered relics' but have been carefully constructed or borrowed according to various patterns. In fact, because of spelling reforms, the actual forms of most words we now use are only a few hundred years old! Furthermore, selecting two of these languages belonging to different families, there is no evidence that there ever was a common origin of German and Hebrew!62
Conclusion
We have seen that in languages there has been a great 'downhill' simplification in inflexions, etc. by natural processes, while the huge 'uphill' growth of languages in their vocabulary and expressiveness has only come about through intelligent human input. These kinds of change are quite different to the processes proposed by biological evolution, so any analogy is completely unfounded.
Acknowledgements
I would like to thank both Dr Noel Weeks and Dr Pieter Honkoop for reading a draft version of this paper and for several helpful suggestions. Allan K. Steel obtained a B.A. (Hons) degree and University Medal in Pure Mathematics from the University of Sydney. He has been a researcher in pure mathematics and computer science for over eight years, in the areas of computer algebra and cryptography. He has a good knowledge of Latin, Greek, Hebrew, German and French, plus a basic knowledge of five other European languages. He has privately studied the history and etymology of several IndoEuropean languages, especially English. Return to top.
Wave model
In historical linguistics, the wave model or wave theory (German Wellentheorie) is a model of language change in which a new language feature (innovation) or a new combination of language features spreads from a central region of origin in continuously weakening concentric circles, similar to the waves created when a stone is thrown into a body of water. The theory was intended as a substitute for the tree model, which did not seem to be able to explain the existence of some characters, especially in the Germanic languages, by descent from a Proto-language.
The tree model requires definite, stable languages, exactly what was denied by the Wave Model; if there are no permanent languages, then they cannot evolve as a tree. Conversely, the Wave Model regards languages as impermanent collections of features at the intersections of multiple circles. What really exists are dialect continua. Johannes Schmidt used a second metaphor to explain the formation of an impermanent language from a continuum. The continuum is at first like a smooth, sloping line. Speakers in close proximity tend to unify their speech, creating a stepped line out of the sloped line. These steps are the dialects. Over the course of time some steps become weak and fall into disuse, while others preempt the entire continuum. As example Schmidt used Standard German, which was defined to conform to some dialects and then spread throughout Germany, replacing in many cases the local dialects. Despite their intent, the wave theorists did not manage to overthrow the Tree Model. Instead it has only been invoked[by whom?] to complement the Tree Model in cases where the latter cannot explain some features. The basic Tree Model has remained unusually effective[clarification needed] and resilient. In modern linguistics, the wave model has contributed greatly to improve the tree model approach of the comparative method.[1]
Diagram based on the Wave model originally presented by Johannes Schmidt. In this Euler diagram, the circles are to be regarded as diachronic; that is, they increase in diameter over time, like the concentric waves on a water surface struck by a stone. The background represents a dialect continuum of no language boundaries. The circles are stable dialects, characters or bundles of characters that have been innovated and have become more stable over an originally small portion of the continuum for sociopolitical reasons. These circles spread from their small centers of maximum effectiveness like waves, becoming less effective[clarification needed] and then dissipating at maximum time and distance from the center. Languages are to be regarded as impermanent sets of speech habits that result from and prevail
in the intersections of the circles. The most conservative language is represented by the area not covered by the circles.