Part 2

The languages of the European peoples are to a large extent grouped into Celtic, Romance, Teutonic, and Slavonic families, all of which have sufficiently similar features to be classed together with Sanskrit as Indo-European or, to use a much-disputed term, Aryan. Asiatic immigrants into Europe both in Arctic regions {17} and on the steppes of South Russia have brought in Asiatic languages which have managed to persist as far west as Lapland, Finland, Esthonia, and Hungary, but they are not sufficient to affect the general statement.

The prior home of the Indo-European languages has been discussed almost _ad nauseam_, and it is still so unsettled a question as to make it unprofitable to discuss afresh whether that home was on the Asiatic steppe, in the mountain and plateau country from Asia Minor to the Hindu Kush, on the grassland borders of South Russia, or somewhere near the Baltic.

It would at any rate seem that these languages did not originate in the Mediterranean, and that the languages of that region in pre-classical times were of other affinities. Minoan, Etruscan, &c., are still undeciphered. If then the classical languages came into that region from outside, the probabilities are that they came in from the north, where are found related languages which have certainly not come from the Mediterranean.

The only language which is of older date in Western Europe than those of the Indo-European family is Basque, still spoken in and near the western Pyrenees. It is a mistake, however, to speak of a Basque race, for the physical types to north and south of the Pyrenees are predominantly broader and longer headed respectively. Educated people in Basque country speak either French or Spanish in addition to the mother tongue, and the Basque language has thus taken up many words from its neighbours; it has very little literature and has contributed extremely little, so far as is known, to French and Spanish. Its affinities are quite unknown, though a kinship with the Bronze Age languages of the Mediterranean is suspected, and suggestions of relations to languages of North Africa have been made.

Of the Indo-European languages it seems to be generally agreed that Lithuanian is the most primitive surviving in Europe, and may be the most primitive of all.

{18}

This might be held to imply that Lithuania is near the ancient home of these languages, but it is more probable that old forms survive here because it is a forest country shut off from the great highways of human movement by the immense Pripet swamps to the south of it. It is as likely a place for a 'refuge of the past' as any spot in Europe. Lettish is related to Lithuanian, but the Letts have long been active and commercial, and their language has shed archaisms and borrowed a good deal from Slavonic tongues. Prussian is now quite extinct. Meillet describes this little group as Baltic; and it is related to the Slavonic family. Other writers give the stricter name of Lettic. The Albanian dialects in the Illyrian Alps are another ancient survival with unknown but doubtless Indo-European affinities, and also a considerable Romance element.

Albanian and Letto-Lithuanian apart, the four great families of Indo-European language, Classical and Romance, Celtic, Teutonic, and Slavonic, must have become distinct before the days of the Roman Empire, and must already at that time have developed diverse forms in their various areas. The Classical languages seem to have reached Greece and Italy from the north; how and when we shall not attempt to speculate. The Celtic languages now spoken by the varied and mixed racial stocks on the westernmost fringe of France, Britain, and Ireland, include two groups, one of which, in its use of the _q_ sound and in other features, is held to be fairly close to the root language of Latin, while the other often uses _p_ instead of _q_.

The q-Celtic languages spoken in parts of Ireland, Isle of Man, and the Highlands of Scotland are usually supposed by archaeologists to have spread westward from Central Europe with the men of the leaf-shaped (bronze) sword about 1000 B.C. The view of Zimmer and Kuno Meyer that the q languages reached Ireland via France without touching Britain is now held to be less probable. The p-Celtic languages are almost universally {19} associated with the invaders of Britain in the pre-Roman Iron Age (the La Tene period). That these men occupied south-east Britain to a considerable extent there can be no doubt, and it is highly probable that their physical type is far more characteristic of the English plain to-day than of any part of Wales, save probably the Severn valley. Yet it is Wales that talks the p-Celtic language, which may have reached it in Roman times, while the Severn valley and the English plain have taken to English speech. Thus we see how inadvisable it is to speak of language-groups as race-groups.

The Baltic or Lettic and the Celtic languages are in essentials survivals of antiquity, of great interest and value as preserving elements which would otherwise have been lost by process of time, but of uncertain future, for no man can say to what extent they are likely to be spoken a few centuries hence. It is otherwise with the Romance, Teutonic, and Slavonic families of languages; they are the definite possessions of the peoples of various regions for centuries to come, and it will be well for us to try to realize how this has come to pass.

Whatever their origins, the Classical tongues spread southward into the Mediterranean. Greek was propagated far and wide by trade and by the movements of which Alexander's march to India was a sign as well as a factor. Rome gathered up into itself the general heritage of antiquity for better and for worse, and associated it with the Latin language, which became a lingua franca in some sort within the bounds of the Roman dominion, so that it influenced deeply not only the language of Gaul but also that of Britain (the p-Celtic language). The power of Latin was so great that only insensibly did the more modern languages, Portuguese, Spanish, Catalan, Italian, Provençal, and French, arise out of it, leaving ancient forms like Walloon, Romansch, Ladin, and Frioul in remoter spots. Their rise was an event of the Middle Ages, though the habit of using Latin for scholarly purposes persisted far longer. In earlier times languages evidently {20} were more fluid. Walloon is practically French with Latin and Celtic elements rather stronger than they are in Parisian, and we have little doubt of the Latinization and subsequent Gallicization of a previously Celtic-speaking people. But since the Middle Ages language frontiers have hardly moved.

This fixation of language has several factors.

In the first place, while villagers beginning to conquer a forest may be feeble and isolated, and liable to rapid change of domination and of culture relations, the process of closer settlement brings them into relation with one another, at any rate over small areas, and gives language a much greater hold on a larger number of people.

In the second place, closer settlement generally implies the evolution of markets, substituted in some cases for the seasonal fairs that were sufficient in times of smaller needs and sparser population. And the market with its settled population of tradesfolk and lawyers makes a language centre. With this goes also some development of communications, or at least of their use, and the possibility of the rise of language beyond mere localism. The crusaders, pilgrims, students, and minstrels of the Middle Ages need to be in the student's mind in this connexion.

In the third place, the development of the art of writing and its increased use are great factors of language fixation, and associated with this is the development of a settled legal system.

We should have these points in mind in trying to understand the evolution of the modern distribution of the languages belonging to the three great families, an evolution which has occurred to a large extent since the fall of the Roman Empire.

{21}

3

_The Peoples of Romance Speech_

Rome may be said to have gathered up the heritages of antiquity and to have passed them on to that part of Europe which the Empire administered, profoundly influencing for a time at least the intimate life, and therefore the language, of the people in all parts in which it found people not as yet accustomed to writing.

In the Aegean, Greek was old established, and intercourse had developed a unified language out of the variants and dialects of earlier times; hence Latin did not impose itself, and a modified Greek has persisted down to modern times, with the serious limitation that the written language is very different from the conversational one, and that therefore the vitality of literature is much reduced, and the life of the people held back thereby.

It has often been noticed that whereas Latin spread far and wide from Rome to Portugal, the Rhine and Rumania, Greek did not spread, or at least did not maintain the majority of its temporary spreads, from the Aegean.

The reasons for this are doubtless numerous, but among them one may note that the spread from the Aegean could hardly be into regions just being opened up for agricultural settlement, and that the Latin spread was associated with the spread of a great scheme of communications and of legal organization. Again, one should realize the localism which was so deep-seated in Greek feeling, and contrast it with the famous idea of 'Civis Romanus'. In accord with this we naturally find that the Eastern Church translated the sacred books into the Slav languages, whereas the Western Church has insisted on Latin in a way that has been full of the most momentous consequences for modern Europe. We must also remember the progressive hemming in of the Greek region by Islam at the time when languages were growing out of {22} Latin in the west, so that the Greek base had nothing like the expansive power of the Latin one.

Facts of physical geography had a good deal to do with the rise of the Romance languages from Latin. The Empire reached to the Danube and Rhine and over the English plain. It had some hold over Wales, and for a time over South Scotland, but none over the regions where Celtic has persisted in Gaelic form. It is abundantly proved that the hold of Rome over Britain was less close than that over Gaul, where much Romanization of the leader classes occurred, and the Romanized urban life was fixed for all later time in the south. Fragments of evidence from Silchester and elsewhere go to show that Latin was widely used in urban life, and so its vocabulary penetrated the language of the pagani or peasantry, and Welsh to this day retains many traces of this contact with Latin. But a country with marked contrast of speech between peasant and citizen is linguistically weak, and the Anglo-Saxons were thus helped to spread their language when they arrived after the Romans lost their grip and many Romano-British leaders had emigrated to Brittany. Romance languages do not now reach the banks of the Rhine anywhere, though they approach them fairly closely at the gate of Belfort, that approach to the Rhine where southern influences would spread most powerfully. Wallony, as a dissected plateau, rather remote and backward until the coal period, has retained its Romance speech, and has increasingly assimilated it to modern French. Lorraine, separated from the Rhine basin by the forested heights of Hunsrück and Hardt, has also remained stubbornly French, but German has spread up the Rhine tributaries through North Switzerland, up the Ill basin, which is now Alsace, up the Moselle gorge between Hunsrück and Eifel, so that Trier (Augusta Trevirorum, Treves) is now German, up the tributaries of the lower Moselle, so that the peasantry of Luxembourg, the people of the Eifel and of the Saar basin, are largely or even mainly {23} German speaking, and along the lowland to Flanders, which maintains Low German dialects, collectively called Flemish. The Flemish plain and the Walloon plateau were long separated by the forests of Hainaut and Brabant, but the clearing of those forests in the Middle Ages, mainly by people spreading down from the plateau, extended the area of Gallic speech and reduced the chances of development of a real linguistic unity in the Flemish fringe. Flanders in consequence suffers from disparity of dialects, and French has been widely used as a language of civilization. With the rise of the nationalist sentiment (see p. 98) in modern times, a severe reaction against the dominance of French has set in, and now threatens the very existence of the Belgian kingdom. A part of the département du Nord, in France, speaks Flemish, but French has long been gaining ground slightly on Flemish in France and probably in Belgium too, in spite of a modern spread of Flemish workpeople into the French speaking area.

East of Switzerland the great Alpine barrier between the Roman bank of the Danube and the real Latin areas brought it to pass that the pagani near the Danube were not fully Romanized, and made the whole region fall linguistically to the invading languages, German and Magyar, but a small island of Ladin, a collective name for several ancient dialects, still persists around Cortina and in the upper Grodenthal. The allied Frioul (with more Italian admixture) is spoken in eastern Venetia, and Romansch, also related, in the south of Grisons, East Switzerland. The main boundary between Romance and Germanic is, as usual, not along a main watershed. The watershed, which is pierced by the Brenner Pass, has German spoken on both sides, and the agency town on its south side (Bozen) is still German in language, though strategic considerations have made the victors move the new bounds of Italy right up to the Brenner.

A curious result of language changes is observable in Rumania. {24} The Roman frontier province of Dacia was strongly occupied for a time (A.D. 107-255), and the Low-Latin element must have had a marked cultural superiority over the indigenous migratory shepherds. They imposed their language, and the Transylvanian mountains have assisted its survival ever since, but it has survived only thanks to large borrowings which have made its vocabulary three-fifths Slav. In spite of military pressure from time to time, the people of this hill knot have followed the usual rule for people so situated, and have spread downhill, giving an area of modified Romance language from the Dniester almost to the Tisa (Theiss) and from Bukovina to the Balkan Danube. This area was, however, seriously isolated from the other Romance regions after A.D. 270, and it naturally received Christianity from Constantinople, so that it has come to be distinct (p. 79) religiously from the other areas of Romance speech, which are the stronghold of the Roman Catholic Church.

The large Slav element in Rumanian well illustrates the general principle that vocabulary is more easily changed than syntax, a principle one may follow along the Flemish border, where Flemish idiom is maintained with French words, or along the Welsh border, where the form of sentence is so often Welsh though the words be English.

Speaking generally we see that there has been on the whole a certain recession of Latin from the old boundaries of the Roman Empire, but that, this apart, that Empire exercised a most potent influence on the speech of its European citizens. The recession was due to mass invasions which, however, as a general rule, penetrated much farther into the Roman domain than one would judge from linguistic evidence alone; in other words, the Latin influence has surged up again, and indeed in one sense those repeated resurgences of Latin feeling have been a main feature of European life in the last twelve or fourteen centuries, while the swaying of the power of Latin and non-Latin elements has {25} been one of the chief causes of war and disunion on the Continent. We may now glance at a few aspects of this swaying of boundaries which are germane to the object of this little book.

The Roman Empire in Gaul seems to have been divisible into belts, the southern of which, with its dry and sunny summer, became deeply Romanized in city and country, language and tradition, a zone in which the cities were the residences of all who had sufficient wealth. The middle one, in the main the Paris basin, had fewer cities, but the country was Romanized in language at least. The northern one near the Rhine had strong Roman cities, especially frontier-towns along the river, but the country does not seem to have been Romanized at all deeply.

The Franks seem to have been bodies of adventurers seeking new homes, the surplus population of Teutonic regions in what is now north-western Germany. They were not trained to city life, and though the strong frontier-cities survived their passage, the cities of the Paris basin seem to have weakened under their onset, albeit the invaders were not able to break the continuity of civic life in the south. The Paris basin thus became provided with a rural Frankish landed aristocracy around which later on there developed the localism which is called the feudal system. As is the way of aristocracies the rude Franks adopted and modified the Romanized language of the Gauls, making the _langue d'oïl_ the mother of modern French. In the south, older forms persisted less altered as the dialects of the _langue d'oc_.

Though in more modern times French has more or less triumphed in the south, two variants of the old languages persist and have acquired some literary strength; they are Provençal, in use in Provence, east of the lower Rhone, and Catalan, in use in Catalonia, and to some extent in Roussillon, modified by contact with Spanish. Needless to say the _patois_ of the peasantry of southern France retains many traces of the _langue d'oc_, and the {26} boundary between the _langue d'oc_ and _langue d'oïl_ has been traced by French scholars. A study of that boundary shows that the _langue d'oïl_ penetrated through the gate of Poitou between the Central Plateau and the barren lands of La Gatine, and established its hold in the basin of the Charente, and almost as far as the gates of Bordeaux. Farther east the lower slopes of the Central Plateau form its effective boundary; the export of men and women downhill from the plateau being apparently sufficient to restrain the general tendency to uphill spread of the language of the plain. In the Rhone valley the boundary of the _langue d'oc_ bends southwards, so that Lyons belongs to the _langue d'oïl_, and the whole Isère region is intermediate between north and south. The boundary near the Rhone is at the narrowing of the river valley near Valence. The narrow, formerly forested, section of the Rhone valley between Valence and Donzère has been a barrier in several ways; Mr. Peake has shown that it divided a Burgundian from a southern culture in the earlier phases of the Bronze Age. It was a factor in the southern boundary of the Burgundian kingdom of Theodoric's time (_c._ A.D. 520), in the boundaries of the Comté de Provence and the Comté du Valentinois, and so on. Donzère is approximately the northern limit of the olive, and the Isère basin has forests that are distinctly non-Mediterranean on its great slopes. As one journeys from the Isère region to that of Drôme and Durance one finds the aspect of the country changing from that of the forest of summer green to that of bare rock masses and a marked tendency to summer brown.

That this boundary is not an isolated phenomenon is well seen by comparing it with the boundary between the France which reintroduced the Roman law (_Droit Écrit_) early in the Middle Ages and the France which preferred to go on judging by customary law (_Droit Coutumier_). This boundary and that of language intertwine, but the influence of big cities as legal centres accounts for some differences. The legal boundary naturally keeps some {28} distance north of the linguistic one near the great southern city of Bordeaux. On the other hand, the influence of valley cities like Clermont-Ferrand helped the northern law to penetrate up the slopes of the Central Plateau, so that the legal boundary is here the more southerly. Lyons, a southern city in so many ways, pushed the legal boundary far to the north of the linguistic one in the Rhone-Saone basin. The boundary line of old-time (pre-1789) Customs Dues, the boundaries of the various Salt Dues (_Gabelles_), of the Governments in existence in 1789, of the general prevalence of the Gothic style in architecture, of the general prevalence of (north) steep-roofed and (south) flat-roofed houses respectively and so on, are all related to the above lines, and can be followed in some detail to work out interesting divergences. The net result of a study of these lines is to show that the French people, apart from the German borderland, consist of two cultural elements, a southern with Roman survivals in cities, language, &c., and a more northern one which is much less Roman. Between them is an intermediate area with mixed allegiance. The study of such a set of boundaries can be given a special value, for it shows us that the linear boundary, however necessary it may be under our outworn system of aggressive states in politics, is an artificiality when applied to the study of peoples. Boundaries are zones, not lines, zones of intermediacy which with better political organization might become interpreters rather than causes of conflict as they have been in the past.

In order to understand the German borderlands better it would be well to use similar methods in studying them, though here we have difficulties due to political prejudice. We have seen that in the Paris basin the rural Franks adopted in modified form the Romanized language of the Gauls, and with this came in the influence of the Roman Church, the heir of the imperial tradition, not only here, but also along the old frontier of the Rhine-Danube. Under this influence civic life grew afresh, and the towns of the {29} Paris basin with their market-places dominated by the cathedral are a most notable feature. As these towns grew, the enlargement of the churches under the influence of the great wave of mediaeval enthusiasm led to the adoption of what is called the Ogival or Gothic style of building, in the Ile-de-France, and soon in the Paris basin generally, as a substitute for a Romanesque style which had been tending to change under pressure of Eastern (Byzantine-Lombardic) influences. With the details of architecture we are not primarily concerned, but the spread of the Gothic over the Paris basin, its failure to oust the Romanesque in the south, its penetration along the Route de S. Jacques (Pilgrim Road to Compostella) to Bordeaux, Bayonne, Burgos, and Léon, its later spread through the Flemish lowland and also across the hills to Metz, Strasbourg, and Basle, are a useful indication of the spread of Gallic feeling. In the Rhine region it spread with difficulty, and the Romanesque is highly characteristic here; beyond the Rhine it spread only with much modification.