6.

The Patterning of Dialect Areas

6.1 Introduction

The construction of a scheme of dialect areas on the basis of localized evidence secured by selective sampling is the immediate objective of area linguistics. The procedure leads from the determination of individual heteroglossic lines to the discovery of bundles of such dividing lines in some sectors of the area surveyed. Counting the number of heteroglosses in the several bundles and evaluating them from the structural point of view lays the foundation for dividing and subdividing the total area. The reliability of a scheme of the dialectal patterning of an area depends in part upon the adequacy of the sample and in part upon the point of view governing the structural evaluation of the heteroglosses forming the boundaries. Any such scheme has the character of an approximation to the linguistic realities and is subject to revision and open to refinements in the light of new evidence or of new insights.

Within the range of purely linguistic operations, a dialect scheme based upon localized evidence functions as a frame of reference. The dissemination of individual features can thus be described in linguistic terms without reference to physiographic, demographic, political, economic, or cultural domains. One can say, for instance, that in the Eastern United States stoop “porch” is confined to the Northern dialect area, lightwood “kindling” to the Southern area, the low vowel / a / in aunt, calf to the subarea of Eastern New England, the checked vowel / U / in coop to an area comprising the South and the South Midland [Kurath 1949: Maps 7, 29; Kurath-McDavid 1961: Maps 67, 68, 108]. Such descriptions are brief, clear, and to the point, whereas premature reference to extralinguistic domains is apt to be misleading.

The construction of a scheme of dialect areas has a further purpose: It is an indispensable preliminary to the sociocultural interpretation of the linguistic situation, both synchronically and diachronically. However imperfect or tentative such a scheme may be, it furnishes the basis for discovering congruences between linguistic and sociocultural domains and their boundaries, which in turn set the stage for more or less probable historical interpretations.

Experience has shown that dialect areas are often roughly congruent with political or ecclesiastic domains, with trade areas and transportation systems, with settlement areas, and even with physiographic provinces insofar as they underly sociocultural domains. Concentric dialect boundaries are frequently focused on important cultural centers from which innovations spread to the receptive hinterland. More often than not, a combination of sociocultural factors is involved in shaping a dialect area. Moreover, with changes in social organization the linguistic patterning of the total area also changes: the past is closely interwoven with the present.

Every language area presents its own problems of historical interpretation. Examples taken from the regional behavior of American English illustrate this point (see Chapter 3). The linguistic “profiles” of several European countries that follow should add substantially to an understanding of current practices in area linguistics and call attention to its unique contribution to linguistic science. They are drawn from an extensive body of scholarly literature of the last fifty years.

6.2 Southern England

How a sampling survey carried out on a modest scale can reveal important aspects of the dialectal structure of an area is strikingly illustrated by Guy S. Lowman’s investigation of southern England thirty years ago (1937-38). Lowman chose sixty communities and interviewed a single speaker of the local folk dialect in each of them. His questionnaire consisted of nearly 400 items, chiefly phonological, chosen from the worksheets of the American linguistic atlas. The intention of this undertaking was to provide background material for the historical interpretation of regional and social differences in American English. This “American” bias in the choice of items cannot, of course, provide a balanced view of the dialectal structure of the area investigated. Nevertheless, the findings furnish important insights, as shown on the accompanying sketch maps, on which the locations of the “listening posts” are indicated by dots.

The four heterophones presented in Figure 25 are highly suggestive with respect to the dialectal structure of the southern part of England. The Home Counties (the London area) are clearly shown to be a part of the East Midland dialect area; they are rather sharply set off from the western area(s). South of the Thames the spacing of the heteroglosses indicates a gradual southward expansion from the London area.

The three heteroglosses presented in Figure 26 are highly suggestive of the relations between the London area and its surroundings.

Line 3, along with the heteroglosses shown in Figure 25, indicates that the London area is part of an extensive eastern dialect area—a well-known fact.

Line 2 suggests a tie-up with the industrial area of the Midlands, which remains to be investigated. Important lines of communication, as the railroad lines leading from London to Birmingham-Manchester and to Nottingham-Leeds, may well be involved as favoring diffusion.

Line 1 sets off an area—the Home Counties—to which urban London usage is apt to spread by “pendular” population movements.

The field records on which the seven heteroglosses presented above are based are part of the collections of the Linguistic Atlas of the United States. Lowman was exceptionally well qualified to carry out this survey for the Linguistic Atlas. He received his doctorate under Daniel Jones in 1930 and spent seven years doing field work for the American atlas before returning to England for a year.

Harold Orton’s Basic Material of the Survey of English Dialects will greatly increase our resources for dealing with the problem so briefly dealt with here.

6.3 Dialect Areas of Medieval England

That a small number of securely established heteroglossic lines can serve as a meaningful framework for suggesting the dialectal structure of an area is also illustrated by the survey of regional traits of Middle English as reflected in localized documents written between ca. 1400 and ca. 1450. Carried out in preparation for editing the Middle English Dictionary, which undertakes to document regional usage by drawing upon appropriate texts, the results were published by S. Moore, S. B. Meech, and H. Whitehall in Middle English Dialect Characteristics and Dialect Boundaries [1935]. Their chief findings are summarized and tentatively interpreted in the Plan and Bibliography of the Middle English Dictionary [Kurath 1954].

FIGURE 25: Southern England: West versus East

1. In the east, postvocalic / r /, as in ear, care, four, has developed into the semivowel / /; in the west it survives as a constricted /r/.

2. In the east, law has a well rounded vowel, [ɔ], in the west an unrounded vowel, [ɑ ~ ɑ].

3. In the east, ME / ā / and / ai, ei /, as in tale and tail, are merged in an upgliding diphthong [ei ~ εi ~ ai]; in the west, ME / ā / appears as an ingliding diphthong [e^ə] and ME / ai, ei / as an upgliding diphthong [εi ~ aei].

4. The syllabic of down, out is articulated as / æu ~ εu / in the east, as / əu ~ ɐu / in the west and in East Anglia.

FIGURE 26: Southern England: London and Environs

1. In London and environs (the “Home Counties”), ME / ō / and / ou / are merged in an upgliding diphthong [ou ~ ɔu], as in stone and grown. Outside this area, they remain distinct as ingliding [o^ə ~ u^ə] ≠ [ou ~ ɔu], respectively.

2. In the Home Counties and in a corridor leading northward, nine and right have a “slow” rising diphthong with a low-central to low-back beginning, i.e. [ɑ^•i ~ ^•i]. Outside of it, a “fast” rising diphthong starting in a centralized position, i.e. /ɐi ~ əi /, is current.

3. In common with the greater part of the eastern counties south of the Wash, the London area has the checked / ∪ / of foot in the word room. To the west, room has the free vowel /u / of do.

Figure 27 presents the six most important, or striking, heteroglossic lines.

North of line 1, OE /ā / as in stān “stone” remained unrounded, to the south it became rounded to open / ǭ /. From sources other than those used by Moore [1935], we know that north of this line OE long / ā / merged with OE short / a- / in open syllables and with the diphthong /ai / from OE /æg /; south of it, OE / ā / merged with OE short /o- / in open syllables. The structural importance of this diagnostic heterogloss is thus quite clear. It sets the Northern dialect area off from the Midland area.

South of line 4 the present plural of verbs ends in -eth, north of it in -en. According to Moore this morphological heterogloss is rather closely paralleled by the line separating the southern voiced initial fricative /v- / from the voiceless /f- / to the north of it, as in foot. From other sources of information we gather or infer that in all probability initial / s / and / Ƿ / have also become voiced within the southern / v- / area. This bundle of heteroglosses can safely be taken to reflect the dividing line between the Southern and the Midland area at ca. 1400.

Line 5 sets off a western area in which by ca. 1400 four rounded front vowels had been preserved: /y /, as in the word “hill”; / ȳ /, as in “fire”; / ö / (< OE /eo /) as in “self”; and / ȫ / (< OE / ēo / ) as in “lief.” To the east of this line, these vowels had merged with unrounded front vowels of the same tongue level. This bundle of structural heteroglosses should surely be recognized as the most important boundary between the West Midland and the East Midland dialects of Middle English, though it was rather sharply recessive. According to A. Brandl and H. C. Wyld, two centuries earlier this bundle of heteroglosses was located much farther east [Kurath 1954: 10].

Line 6 has no structural implications. Separating Western short / o / before nasals from Eastern / a /, as in the word “man,” it is merely a convenient diagnostic feature.

Line 3 separates the variant inflection of the present singular 3: -es to the north vs. -eth to the south. In its eastern sector this line is paralleled by the heteroglosses them (< Old Norse) ≠ hem ~ hom (< OE) and sal ≠ shal. This bundle of lines subdivided the East Midland area into a northern and a southern section, perhaps as late as ca. 1450. By ca. 1500, North Midland -es and them had been carried southward to the London area to become established in the literary language.

Other regional features of Middle English have as yet not been delimited with any precision. For the time being, their location can be best described with reference to the skeletal structure suggested by well known heteroglosses. This, indeed, is one of the functions of a set of diagnostic heteroglosses.

FIGURE 27: Middle English Heteroglosses.

1. The vowel in “stone”: N / ā / ≠ S / ǭ / (OE ā)

2. The inflection of the pres. pl.: N –es ≠ S -en

3. The inflection of the pres. sing. 3: N –es ≠ S -eth

4. The inflection of the pres. pl.: N –en ≠ S –eth

5. The vowels in “hill, fire, self, lief”: W rounded /y, ȳ, ö, ȫ/ ≠ E unrounded /i, ī, e, ẹ̄ / (OE y, ȳ, eo, ēo)

It will be interesting to see how the extensive investigation of Middle English localized documents and of localizable texts by Angus Mcintosh [1963] will round out and refine our conception of the dialectal structure of Medieval England.

6.4 The Swabian Dialect Area

An important contribution to the method of delimiting speech areas and of showing their relations to adjoining areas was made by K. Bohnenberger. Formulated after some 30 years of field experience in connection with his fundamental article on the eastern boundary of the Alemannic speech area [1928], he applied it in his posthumously published book on the Alemannic dialects [1953].

The central feature of his procedure is the choice of a representative heterogloss sector by sector. This devise enables him to describe for each sector the course of the several heteroglosses with reference to it, pointing out whether they coincide with it or deviate from it more or less. This, in turn, leads to the description of dialect boundaries as sharply defined or as transition belts. Finally, the character of any given dialect boundary, sharp or graded, furnishes valuable leads to its interpretation in sociocultural terms.

For the Alemannic area Bohnenberger had at his disposal about three dozen heteroglosses whose location he had ascertained very largely by his own fieldwork. These he assembled on a single map (11 by 13 inches), identifying the heteroglosses and their Middle High German sources [1953: 301-2]. Up to this point his procedure is strictly synchronic. The map exhibits the geographic behavior of the heteroglosses as of ca. 1900-1950. It shows that the Swabian subarea of Alemannic is set off by rather striking bundles of lines, close-knit in some sectors and spaced in others.

At this stage, Bohnenberger selects specific heteroglosses to represent the bundles sector by sector. Since there is no one heterogloss that runs through all the bundles that frame the Swabian dialect area—no Swabian shibboleth comparable with Bavarian enk “you” (plural)—this procedure imposes itself. On his map the author draws these representative or diagnostic heteroglosses as heavy lines, a rather effective cartographic device.

Bohnenberger’s choice of diagnostic heteroglosses is diachronically oriented. His intention is to prepare the ground for a historical interpretation of the linguistic situation obtaining at the turn of the nineteenth century. The author is fully aware of the fact that his choice of diagnostic heteroglosses is bound to be somewhat arbitrary and that “mixing” the diachronie with the synchronic point of view has its hazards. But he is convinced that there is no better way of dealing with the situation. In any event, his procedure provides a meaningful perspective of the Swabian dialect area in its relation to the other Alemannic areas as well as to the adjoining Bavarian and Frankish domains.

The accompanying sketch map (Figure 28) presents the diagnostic heteroglosses chosen by Bohnenberger to “frame” the Swabian dialect area and to show its relations to the adjoining speech areas. Brief comments on their character are appended.

Line 1 separates the Swabian from the Bavarian dialect area. It represents a bundle of heteroglosses that follow the Lech River from the Alps to the Danube; close-knit south and north, they are spaced along the middle course of the Lech.

Line 2 is the easternmost heterogloss in a widely spaced bundle that forms the transition from Swabian to Frankish.

Line 3 separates Swabian from Frankish. East of the Neckar River it is closely paralleled by other heteroglosses; farther west it marks the southernmost extent of a transition belt.

Line 4 separates Swabian from Low Alemannic. It represents a bundle of heteroglosses that run close together along the eastern slope of the Black Forest and again north of Lake Constance, but are more or less widely spaced elsewhere.

Line 5 serves to set Low Alemannic off from Frankish. It marks the northern extent of a transition belt.

Line 6 suggests the boundary between Low Alemannic and High Alemannic.

6.5 The German Area

The earliest dialect scheme of an extensive area based upon heteroglosses established by a systematic selective survey is that of F. Wrede, director of the Sprachatlas des Deutschen Reiches for more than 20 years. It was published posthumously as Map 56 of the Sprachatlas [1934].

As early as 1919 Wrede advocated the use of diagnostic indicators (Merkmale) on the model of the satam ≠ centum heterogloss in comparative Indo-European linguistics in these words: “This kind of device is clear, whereas, from the point of view of area linguistics, the traditional distinction between High German and Low German is ambiguous” [Wrede 1963 < 1919: 340].

F. Wrede’s plan of subdividing the German speech area on the basis of individual diagnostic heteroglosses is presented in Figure 29 on a simulated map.

It will be observed that Wrede accepts the traditional nineteenth century grouping of the German dialects into Low and High and the subdivision of High German into Central and Upper. Equally traditional is his subdivision of Low German into (a) Low Frankish, (b) Low Saxon, and (c) East Low German; of Central German into (a) Western and (b) Eastern; and of Upper German into (a) Alemannic and (b) Bavarian.

FIGURE 28: The Swabian Dialect Area

Shown in its relation to the Low Alemannic (LA), the High Alemannic (HA), the Bavarian (B), and the Frankish (F) domains.

Heteroglossic Lines

1. Swabian reflexes of MHG ir, iuch “you” (PGc pi.) ≠ Bavarian es, enk “you” (PGc dual)

2. Swabian / ā /, as in / dāg / “day” ≠ Frankish / ö / (OHG ă)

3. Swabian / ǫa ~ oi /, as in / broad / “broad” ≠ Frankish / ai ~ ā / (OHG ai - ei)

4. Swabian / ei /, as in / eis / “ice,” and / ou /, as in /hous / “house” ≠ Low Alemannic /ī/ and / ū / (OHG ī and ū)

5. Low Alemannic / pf- /, as in / pfund / “pound” ≠ Frankish /p- /

6. Low Alemannic /kh- /, as in /khopf / “head” ≠ High Allemannic /xopf /

Based on K. Bohnenberger, Die alemannischen Mundarten

What, then, is Wrede’s contribution to the method of subdividing the German dialects? It consists in basing each boundary, major or minor, upon a single heterogloss whose course had been established by the Sprachatlas. Thus the boundary between the Low and the High German areas rests upon the heterophone /-k ≠ -x / as reflected in ik ≠ ich; that between the Alemannic and the Bavarian areas upon the heterolex euch ≠ enk “you” (pl. ); that between the Low Frankish and the Low Saxon areas upon the heteromorph -en ≠ -et (inflection of the present plural).

The scientific as well as the practical importance of this step is clear. First of all, it furnishes the investigator with guidelines for discovering and describing heteroglossic lines of similar trend on which dialect boundaries can be established. Moreover, it commits him to a consistent orientation, so essential in dealing with complicated linguistic situations.

FIGURE 29: The Dialectal Structure of the German Area

According to F. Wrede

Wrede’s plan fulfills the limited objectives for which, indeed, he intended it. Much of what has been written by Wrede’s students and by other scholars associated with the Sprachatlas group betrays his orientation, for good or ill.

Serving primarily the practical purpose of orientation, the diagnostic heteroglosses (Merkmale) chosen by Wrede involve no differences in phonological or morphological structure between the German dialects. The heteroglosses ik ≠ ich, appel ≠ apfel, and pund ≠ fund exhibit only differences in the incidence of shared phonemes. The -en ≠ -et morphemes are grammatically synonymous, denoting plurality in the present tense of the verb. The euch ≠ enk heterogloss is lexical: no grammatical difference is involved, since both variants function as plurals.

Thus Wrede’s diagnostic heteroglosses are trivial in themselves. The significance they have for a scientific grouping of the German dialects depends entirely upon their presumed representative character, that is, upon the extent to which they are supported by bundles of heteroglosses running similar courses. It can be assumed that, owing to his long preoccupation with the collections of the Sprachatlas, the author had a “feel” for their relevance; and some of his “hunches” have been amply documented by later research.

No attempt has been made in later years to group the German dialects on a different basis. As a matter of fact, it will not be possible to establish a sound classification until at least the major dialect types of German shall have been analyzed from the structural point of view, so that the systemic phonological and morphological heteroglosses can be distinguished from the incidental. Moreover, the areal grouping of morphological and lexical phenomena will have to be taken into consideration.

6.6 The Netherlands

A conspectus of the linguistic situation in the Low Countries—the Netherlands and northern Belgium—is offered by A. Weijnen in De Nederlandse Dialecten [1941]. Assembling 35 phonological and 8 morphological heteroglosses for which fairly adequate data had been recorded by various scholars (including E. Blancquaert, L. Grootaers, and G. G. Kloeke), he presents them on a single map [174] without evaluating them structurally or weighing them from the diachronie point of view. Lexical heteroglosses are deliberately set aside. The result is a strictly synchronic picture, highly suggestive, though admittedly incomplete and tentative.

On this evidence, a relatively uniform central area extends all the way from the French (Walloon) language boundary northward to the Zuider Zee (from Brussels and Antwerp to Amsterdam). In the southwest, Flanders and Zeeland, especially the latter, are set off from this central area by prominent bundles of heteroglosses. The southeast (Limburg) has the character of a transition area. Less prominent bundles subdivide the northeastern Netherlands and set them off from the central area. North Holland and Friesland form another subdivision of the Netherlandish language area.

Whether the large central area is actually as uniform as it appears on Weijnen’s map is subject to doubt. It seems likely that the “deviant” usage of the peripheral areas attracted the attention of scholars and that dialects of the central area, which are closer to Standard Dutch, had been less thoroughly investigated by 1941. At any rate, Kloeke [1950] published a number of phonological and lexical maps that exhibit a west-to-east gradation between the Zuider Zee and “the Rivers.” Figure 30 presents some of Kloeke’s data in simplified form.

If we apply Bohnenberger’s method of selecting diachronically diagnostic heteroglosses to Weijnen’s synchronic map, so as to prepare the way for the diachronie interpretation of the synchronic data, we might choose the heteroglosses presented in Figure 31. As a matter of fact, Dutch and Flemish scholars, preoccupied with historical problems (time depth) as they are, have used such heteroglosses, but often without critical evaluation or proper perspective.

Line 1 sets off the area in which the syllabic of muis “mouse” is an upgliding diphthong / öü /; to the east and in the southwest monophthongal /ǖ/ (< PGc ū) is current.

Line 2 sets off the coastal area in which PGc / 0 / escapes palatal mutation, as in groen / xrun / “green. “

Line 3 separates the northwestern region (North Holland and Friesland) where PGc / ä / was shifted to a front vowel, mid or high, as in Anglo-Frisian: thus slēpen “sleep” ≠ slapen.

Line 4 sets off a region adjoining the Low Saxon area of Germany (the greater part of Overijssel and Drenthe) in which the three persons of the present plural end in -t (as in Low Saxon), whereas all other dialects of northern Netherlandish have -e(n).

Line 5 delimits a region adjoining the German dialect area of Cologne (Köln): here Netherlandish ik “I” is confronted by the ich of the provinces of Limburg, which share this feature with the Cologne area.

All of these diachronically diagnostic heteroglosses are embedded to some extent in bundles of heteroglosses exhibited in present day Netherlandish. Employed with proper reserve, i.e., without plunging headlong into the remote past, they contribute a convenient and significant frame of reference for describing the course of other heteroglosses of the present and for tracing historical developments.

6.7 The Dialectal Structure Of Italy

The Linguistic Atlas of Italy [Jaberg-Jud 1928-40] has furnished the localized linguistic data for delimiting the major dialect areas of Italy and their subdivisions. As early as 1937, G. Rohlfs, who had investigated seventy communities in southern Italy for this atlas, drew upon these data to outline the dialectal structure of Italy and to point out the chief sociocultural forces that brought it about [Rohlfs 1952 < 1937: 89-107]. Some of his findings are summarized in Figure 32.

FIGURE 30: The Central Netherlands

1. West ladder ≠ East leer “ladder” [Kloeke 1950: 159]

2. West butter ~ bōter ≠ East botter “butter” [79]

3. West hiel ≠ East hak “heel” [149]

4. West kaas ≠ East kees ~ keis “cheese” [82]

5. West gras ≠ East gres “grass” [173]

Cities: Ad = Amsterdam, Af = Amersfoort, D = Dordrecht, G = Goringhem, H = Haarlem, L ־־ Leiden, R = Rotterdam, T - Tiel, U = Utrecht

The most prominent dialect boundary of Italy follows the crest of the Apennines, separating Northern Italy from Tuscany. The following North Italian (Gallo-Italic) phonological features have their southern limit here:

FIGURE 31: The Dialectal Structure of the Netherlands

Cities: A(msterdam, An(twerp, B(russels, Br(ugge, C(ologne, D(eventer, G(roningen, N(ijmegen, R(otterdam, U(trecht.

Based on A. Weijnen, De Nederlandse Dialecten

(1) The voicing of intervocalic voiceless stops, as in fradel, formiga, pever ≠ Tuscan fratello “brother,” formica “ant,” pepe “pepper”;

(2) the simplification of geminate consonants, as in spala, gata ≠ Tuscan spalla “shoulder,” gatta “cat”;

(3) the loss of final vowels, as in an, sal ≠ Tuscan anno “year,” sale “salt”;

(4) the loss of unaccented medial vowels, as in slar, mdor ≠ Tuscan sellaio “saddler,” mietitore “mower”;

FIGURE 32: The Three Major Dialect Areas of Italy

The boundaries of 7 Northern and of 11 Southern features are shown on the map.

Adapted from G. Rohlfs, La struttura linguistica dell*Italia

(5) the nasalization of vowels before word final nasal clusters, as in p ã ≠ Tuscan pane “bread”;

(6) some lexical items, as incö ≠ Tuscan oggi “today,” orp ≠ Tuscan cieco “blind.”

This linguistic boundary rests on a natural barrier, the crest of the Apennines. But in ancient times this line was an ethnic boundary between the Gaulish tribes and the Etruscans. Moreover, for many centuries it was also an ecclesiastical boundary dividing the archdiocese of Ravenna from that of Rome.

Another important linguistic boundary sets Central Italy off from the South. Here, however, the line of demarcation is not clearcut. It is rather a transition belt formed by a large bundle of more or less congruent lines. The greater part of this belt extends from the environs of Ancona on the Adriatic across the Apennines to the Tyrrhenian Sea south of Rome. The general course of this graded boundary agrees roughly with the northern boundary of the ancient duchy of Spoleto. Even in the early Middle Ages this duchy was separated from Tuscany by a corridor, the Papal States, which provided communication between the Patrimonium Petri and the exarchate of Ravenna. Thus, the natural relations between the South and Tuscany were breached in the Umbrian area. This had a decisive influence upon the linguistic physiognomy of this part of Italy.

The following South Italian features have their northern limits in this transition belt:

(1) The voicing of voiceless stops after nasals, as in mondone “montone, ram,” tiembre “tempo, time,” angora “ancora, again”;

(2) the mutation of accented vowels before final -i, -u as in acitu “aceto, vinegar,” sulu “solo, only”;

(3) such lexical items as frate ≠ fratello “brother,” fémmina ≠ donna “woman, wife,” ferraru ≠ fabbro “smith,” tenere ≠ avere “have.”

The three major dialect areas of Italy can easily be subdivided. In the North, for instance, the boundaries between the dialects of the Piedmont (Turin), of Liguria (Genua), of Lombardy (Milan), and of Venetia (Venice) are quite sharp.

In Italy, Tuscany has been the most important cultural center since the Middle Ages. As the prototype of the literary language, this region has had extraordinary importance in the diffusion of the cultivated vernacular (volgare illustre) to all neighboring regions.

The currency of Tuscanisms throughout Venetia is well known. Even in the Middle Ages Tuscan infusion brought about a profound transformation of the local dialects of that region. Of Tuscan origin are, for instance: (1) the generalization of the diphthong ie, corresponding almost exactly to the Tuscan situation, as in fiele “bile,” piè “foot,” miele “honey”; (2) words like giorno “day,” semola “bran,” donnola “weasel,” grembiale “apron.” This has had the effect of leveling the dialects of the Venetian plain to a much larger extent than those of the other subareas of Northern Italy.

Tuscan diffusion into Lombardy has been more limited. Under the powerful leadership of Milan, this region largely maintained its linguistic independence. For instance, Lombardy is almost the only section of Northern Italy where ancient caput (Lomb. co) has not yet yielded to testa “head.” However, in recent times the influence of the literary language, which is essentially Tuscan, is making itself felt in Lombardy. This infiltration is directed at the heart of Northern Italy (Alta Italia). Little by little, Milan has become an important center of diffusion for the national language. From the capital of the valley of the Po emanate the linguistic influences that have been upsetting the old idioms, gradually pushing them back to the periphery. Thus, Tuscan zio “uncle” has replaced indigenous barba, zia “aunt” has taken the place of ameda (Tecino anda), giovedi “Thursday” has supplanted the earlier giobia. Instead of vener one now says venerdi “Friday,” instead of Lombard calegar one hears only calsular “shoemaker.” In Liguria and in Piedmont, Tuscan natale “Christmas day” has replaced indigenous nadal.

On the dialect of Rome (romanesco) Tuscan influence has been profound. From Rome northward, the greater part of Latium has been exposed to Tuscan penetration for centuries. Since the fourteenth century, the old vernacular of Rome has suffered constant disintegration. The old Southern vocalism has given way to the Tuscan. One no longer says tiempo but tempo “time,” not cuorpo but corpo “body,” not pede but piede “foot,” not ditto but detto “said,” not munno but monno “mondo, world.” The modern dialect of Rome is simply Tuscan superimposed on the relics of an indigenous substratum.

Various languages have left their marks in the vocabulary of different sections of Italy. North of the Apennines some Celtic words survive, and Germanic invaders—Goths and Lombards—have contributed to the regional vocabulary. In southern Italy and Sicily, French words survive from the days of Norman rule. Other French words spread into Italian by cultural diffusion during the age of chivalry. For all of these lexical admixtures the linguistic atlas of Italy provides ample localized evidence.

The only significant phonological feature of foreign origin admitted by G. Rohlfs is the voicing or “weakening” of medial Latin p, t, k in Northern Italy. This he attributes, with many other scholars, to the Celtic substratum [Rohlfs 1952 < 1930: 75-78].

For general orientation concerning the dialectal structure of present day Italy and its sociocultural and ethnic background, the reader should consult the amply documented and clearly written treatment of this subject by E. Pulgram in The Tongues of Italy (1958), esp. pp. 45–53. In this book the author also discusses the substratum problem with special reference to Italy [esp. pp. 153–54, 203–44, 281, 337–43], in which views diverging from those advanced or implied in the papers by Rohlfs and Hall are duly recorded. See also Pulgram (1949), especially the extensive bibliography and the author’s critical evaluation of divergent opinions from a clearly formulated point of view.

6.8 The Dialectal Structure of France

6.81 The Northern Area

J. Gilliéron’s rejection of the regularity of phonemic change [1919] and his denial of the existence of dialect boundaries—hence also of scientifically definable dialects—did not go unchallenged inside and outside of France. Moreover, essential problems in dialectology ignored or set aside by the author of the Linguistic Atlas of France [1902-10] proved to be a challenge to scholars in the field of Romanic linguistics and philology, among them Karl Ettmayer [1924] of the University of Vienna.

From the first 284 maps of the LAF, Ettmayer chose ninetyeight lexical items that exhibited rather clear regional patterning of variants in northeastern France and adjoining parts of Belgium. After drawing heteroglossic lines item by item, he assembled them all on a single map [Map I]. The result was most striking. He found that the northeastern periphery, extending from Picardy through Belgium to Lorraine, was highly diversified in word usage, whereas the central area, consisting of the Ile de France, Orleans and the Champagne, was relatively uniform. This had, of course, been known in a general way; but the cartographic presentation of the specific data furnished by the LAF dramatized the situation (see Figure 33).

The most prominent bundles of heterolexes, as that between the Ile de France and Picardy or that between the Champagne and Belgium to the north and Lorraine to the east, are labeled with reference to historic political domains. This device provides a terminology for describing the location and course of any single heterolexic line with reference to these bundles. It also suggests, but does not explicitly assert, historical connections between linguistic and political boundaries.

To show lines of diffusion from the Ile de France, Ettmayer selects six lexical items: (1) beau garçon “son in law,” (2) journal “newspaper,” (3) fouet “whip,” (4) bas “drawers, breeches,” (5) baril “cask, barrel,” (6) soir “evening.” Three routes of expansion are clearly brought out: (a) northwestward down the Seine (Rouen), whence in part along the coast to Boulogne; (b) northeastward through the Champagne to Laon-Reims and beyond; (c) eastward through the Champagne in the direction of Nancy. In this process a variety of regional expressions were being replaced by their Parisian counterparts at the turn of the century.

FIGURE 33: Lexical Diffusion from the Ile de France

Lexical Isoglosses

1. beau garçon “son in law”

2. journal “newspaper”

3. fouet “whip”

4. bas “drawers, breeches”

5. baril “cask, barrel”

6. soir “evening”

Cities: B(oulogne, D(ieppe, La(on, Li(lle, N(amur, Na(ncy, Or(leans, P(aris, Re(ims, Ro(uen, T(royes.

Adapted from K. Ettmayer, Über das Wesen der Dialektbildung

Ettmayer raises the eminently pertinent question to what extent morphological and lexical dialect boundaries agree with phonological boundaries. His findings are summarized on his ingeniously constructed Map VII, which shows that there is partial agreement between them in southern France, as south of Lyon and along the Garonne; but there are some sharp divergences elsewhere [44]. This attempt may be premature and overbold; but this problem must ultimately be faced in area linguistics. It calls attention to the partially divergent means by which phonological, morphological, and lexical features can spread.

Written almost as soon as the Linguistic Atlas of France was in print and published, Ettmayer’s essay on the nature of dialect development is still worth careful reading. His perspective is sound and he boldly states his views concerning essential problems that can and must be faced in area linguistics, such as the following: the pervasive influence of dominant cultural centers and of the literary language [11]; the bundling of heteroglossic lines and their complex sociocultural background [13–14]; the expansive effects of communication routes [19]; the retarding effects of regional language loyalty on diffusion [31, 33]; the problem of distinguishing parallel regional shifts from importations [49]; the essential regularity of phonemic change despite obvious deviations from the norm [54].

6.82 The Transition Belt between Francien and Provençal

Relying upon the ample evidence furnished by Gilliéron’s Linguistic Atlas of France [1902–10], R. A. Hall, Jr. [1949], undertook to determine the position of the traditionally recognized Franco-Provençal dialect area within the total linguistic structure of France. For this purpose he chose the eight striking phonological innovations of Northern French (Francien) listed below:

(1) Fronting of accented Latin ā to e after palatals, as in L càrus > cher / šεr / “dear”;

(2) fronting of accented L ā to e after nonpalatals, as in L màter > mère / mεr / “mother”;

(3) the shift of L en to an, as in L centum > cent / s ã / “hundred”;

(4) palatalization of Lc/k / before a, as in L campus > champ / sä / “field”;

(5) palatalization of L g before a, as in L purgāre > purger / pyrže / “cleanse”;

(6) the shift of Romanic medial palatal / λ / to / j /, as in feuille /fOj /”leaf”;

(7) the loss of preconsonantal s, as in L castellum > château /šato/ “castle”;

(8) the loss of medial d (< L t), as in L catēna > chaine / šεn / “chain.”

For each of these features Hall chose from seven to fourteen examples exhibiting a fairly clear regional dissemination of the variants and drew separate heteroglossic lines for them. This cartographic procedure shows at a glance whether in any given locality or district these Northern phonological innovations were adopted in all relevant words or only in some of them. The spacing of the heterophones on Hall’s Maps 3, 4, and 7 demonstrates the fact that they are introduced word by word, i.e., at different times.

When all of the heterophones established on the evidence of the Linguistic Atlas for the eight features investigated are assembled on a single map, Central France appears as a transition zone between the strikingly uniform Northern (Francien) and the relatively uniform Southern (Provençal) areas. The traditionally recognized Franco-Provençal area, Hall concludes, is nothing more than the eastern section of this Central transition belt.

Hall’s carefully documented study not only assigns a meaningful place to Franco-Provençal in the dialectal structure of France. It presents historically important phonological features for subdividing France into major dialect areas and of characterizing each subdivision in phonological terms. The map presented in Figure 34 below, in which the regional dissemination of five of Hall’s eight features have been assembled in simplified form, shows the varying width of the central transition zone between the North and the South.

The complexities of the central transition belt that separates southern from northern France is also effectively conveyed by the compilation of numerous heterophones presented by Ettmayer [1924] on Map VI.

6.9 The Dialectal Structure of the Iberian Peninsula

The Iberian Peninsula harbors three literary languages: Spanish, Portuguese, and Catalan. Each of them is based upon a major regional dialect.

Literary Catalan is used chiefly in Catalonia, although the Catalan dialect is spoken also in large parts of Valencia. Spanish has been the official language since the fifteenth century.

FIGURE 34: France: Outer Limits of Northern Phonological Features

1. Northern e (as in cher) ≠ Southern a (< L cārus)

2. N e (as in mère) ≠ S a (< L māter)

3. N ča- (> š, as in champ) ≠ S ka- (< L campus)

4. N loss of -d- (as in chaine) ≠ S -d- (< L catēna)

5. N loss of preconsonantal s (as in château) ≠ s preservation of s (< L castellum)

Cities: B(ordeaux, G(eneva, L(yon, Marseille, N(antes, Ne(vers, P(aris, R(ouen, T(oulouse, To(urs

From R. A. Hall, Jr., The Linguistic Position of Franco-Provençal.

Portuguese is the literary and the official language of Portugal. It is based upon a group of closely related dialects spoken along the entire extent of the western littoral of the Iberian Peninsula, including Galicia (part of the political domain of Spain since 1230, where Spanish is the official language).

Spanish is the literary and the official language of the entire Iberian Peninsula, except for Portugal. In Catalonia it is rivaled by literary Catalan. Spanish is Castilian in origin and was so called until the rise of Spain as the dominant political power and cultural center of the Peninsula. Castilian is the appropriate term in linguistic discussions.

The linguistic history of the Iberian Peninsula has been extensively investigated since the beginning of the present century. In his Manual de gramâtica historica espanola [1905, 1952] and his Origines del espanol [1926, 1950], R. Menendez Pidal laid the foundation for fruitful research carried on by numerous scholars inspired by him. A great deal of systematic sampling of the modern dialects has also been done, but the Spanish Civil War (1936-39) has seriously interfered with its progress.

A Linguistic Atlas of Catalonia was organized by A. Griera i Gaja and published in part [1923–39]—858 maps. The remainder of his collection was destroyed during the Civil War. T. Navarro Tomas directed a field survey of Spain before the Civil War; unfortunately, none of his findings have been published as yet. In Portugal, M. de Paiva Boléo has energetically promoted the investigation of the dialects since the 1940’s. Reports on the progress of these surveys and their character are available in Kuhn 1947, Pop 1950, Griera 1958.

Although the collections of the linguistic atlas of Spain and Portugal are presumably accessible to qualified scholars, it is regrettable that the Iberian Peninsula does not yet possess a linguistic survey, comparable with that of France and Italy, that scholars in other countries could consult. Nevertheless, the regional dissemination of some phonological features are known well enough to provide insight into some aspects of the dialectal patterning of the Peninsula. The brief sketch that follows relies largely upon the excellent and circumspect treatment presented by A. Zamora Vicente in his Dialectologia Espanola [1960], which also contains a comprehensive bibliography.

Three phonological heteroglosses of structural and historical importance serve to suggest the present dialectal structure of the Iberian Peninsula (see Figure 35). In the large Castilian central area (1) Latin short e and o appear as the crescendo diphthongs ie and uo, as in miel “honey” (< L mel), piedra “stone” (< L petra), and in nuevo “nine” (< L novem), hueso “bone” (< L ossum). (2) Latin initial f- has shifted to h- and is now lost in the greater part of the area, as in Mel “gall” (< L fel), hoja “leaf” (< L folia).

FIGURE 35: Iberian Peninsula: Major Dialect Areas

1. Latin short e and o, as in mel “honey” and novem “nine,” preserved in Gallego-Portuguese along the Atlantic and in Catalan along the Mediterranean littoral, are shifted to ie and ue in the large Castilian central area.

2. Initial Latin f-, as in fel “gall,” folia “leaves,” is preserved in Gallego-Portuguese and in Catalan as well as in northern Le one se and Aragonese, but shifted to h- in Castilian (and thereafter lost, except in western Andalusia).

Cities: B(arcelona, C(ordova, L(isbon, M(adrid, P(orto, S(eville, T(oledo, Z(aragossa.

Adapted from R. Menéndez Pidal, Manual de Gramatica Historica Espanola and from A. Zamora Vicente, Dialectologia Espanola.

In the western and the eastern littorals of the Peninsula, i.e., in Gallego-Portuguese and in Catalan, Latin e, o, and fare preserved as such.

Other phonological features exhibit dissemination patterns that support the present subdivision of the Peninsula into a western (Gallego-Portuguese), a central (Castilian), and an eastern (Catalan) dialect area. Some of these are presented in the accompanying table, to which the numbers used below refer. Features 1 and 2 are confined to Castilian, features 3, 4, 5 to Portuguese, and feature 6 appears primarily in Catalan. In feature 6, each of the three dialects exhibits its own reflexes. It must be pointed out, however, that the precise regional dissemination of some of these features has not yet been established and that diachronie considerations are here not taken into account.

It is a striking fact that the present dialect boundary between Portuguese and Castilian intersects the river valleys and the mountain ranges of the peninsula at right angles. Since there is no congruence between these two dialect areas and the natural landscape, their origin must be sought in sociocultural factors. To some extent this is also true of the dialect boundary between Castilian (Spanish) and Catalan. Spanish scholars, from Menéndez Pidal [1926] to the present, have worked on this problem and achieved substantial clarification of the dynamics that have created the three major dialect areas of the Iberian Peninsula.

The decisive historical events are: (1) the conquest of all but the northern section of the Iberian Peninsula by the Arabs (Moors) in the first decades of the eighth century; (2) the gradual Reconquest of the entire Peninsula (ca. 900-1492) by speakers of Romanic dialects that had developed in the north of the Peninsula along the Bay of Biscay and the Pyrenees; (3) the lateral expansion of the Castilian dialect area after the Reconquest, owing to the rise of Castile as the dominant political and cultural power on the Peninsula.

It is important to point out that in the area dominated for centuries by the Arabs, Romanic continued to be spoken by sizable groups, both in the cities and in the countryside. It remained the language of the Christian congregations. In the millennium that elapsed between the Roman colonization of the Peninsula (200 B.C.) and the Re conquest, regional differences must have developed, especially in areas as far apart as the valley of the Ebro in the northeast, the valley of the Guadalquivir in the south, and the Atlantic coast in the west. Romanic words, found chiefly in Arabic texts, provide some evidence of variation in Mozarabic, as the Romanic language current in the area dominated by the Arabs was called. Such regional differences might well have contributed to the character of the dialects that emerged as the result of the Reconquest.

The Reconquest consisted of a series of southward thrusts from the northern Romanic kingdoms and principalities strung out over a thousand miles from the Atlantic to the Mediterranean, from Galicia to Catalonia. The most vigorous and extensive of these drives was that from Old Castil (Burgos) through the central section of the Peninsula. In the course of about three centuries this southward expansion reached the Atlantic at the mouth of the Guadalquivir in Andalusia. Crossing the Sierra de Guadarrama into the valley of the Tagus, Castile conquered Toledo in 1085, which became the capital of Castil (later of Spain); Seville was taken from the Arabs in 1248, Granada not until 1492.

To the west, the Castilian thrust was flanked by a southward drive of the kingdom of Leon into the valleys of the Tagus and the Guadiana (Estremadura); to the east of it, the kingdom of Aragon expanded southward. On the Atlantic littoral, the Arabs were pushed back from Galicia and northern Portugal, along the eastern shore of the Peninsula from Catalonia (part of the Frankish March maintained by the Carolingian emperors to contain the Arabs of the Ebro valley). These five parallel thrusts created five major dialect areas: the Gallego-Portuguese, the Leonese, the Castilian, the Aragonese, and the Catalan. Each of these dialects is essentially a descendant of a Romanic dialect that survived in the north of the Peninsula, as modified by the regional Mozarabic underlay in the several reconquered areas.

With the rise of Castil as the dominant political power, especially after its union with Aragon (1479), and owing to its cultural predominance from the sixteenth century onward, the Castilian dialect spread eastward into Aragonese and westward into Leonese territory. Moreover, with the acquisition of all parts of the Peninsula, with the exception of Portugal, the Castilian literary language—now called Spanish—became the official language of all of Spain.

The northern section of the Iberian Peninsula has been remarkably conservative in speech. Here Leonese and Aragonese have escaped Castilianization to a considerable extent, as in the triangles between heteroglosses 1 and 2 shown in Figure 35.

All varieties of Spanish spoken on the American Continent, from the West Indies to Central and to South America, are derived from the Castilian dialects of the Iberian Peninsula. Castilian was far from uniform at the time the several American colonies were established. In fact, several important phonological changes were still in progress even in literary Castilian during the sixteenth century, as the merging (ca. 1550) of the earlier voiced fricatives /ð, z, g, / with voiceless /θ, s, x /, respectively, with the result that / θ / became current not only in plaza (< Lat. platea) but also in haçer “make” (< Lat. facere); /s / not only in pasar “pass” (cp. Lat. pass-us) but also in casa “house” (< Lat. casa); and /x / not only in nexo (< Lat. nexus) but also in hijo “son” (< Lat. filium).

While the merging of these three voiced fricatives with their voiceless counterparts and the further merging of / θ / with/ s / in southern Castilian (i.e., in Andalusia)—as in haçer “make” and casa “house”—became fully established in all parts of Spanish America, presumably through continued contacts with the mother country, some other phonological features of Castilian that varied regionally or by social levels still vary in American Spanish. Thus the shift of palatal / λ / to the palatal semivowel / j /, as in llano “level,” lluvio “rain,” became established in the greater part of Spanish America, but not in the Andean highlands from southern Columbia to northern Chile [Zamora 1960: Map XXI].

Zamora, relying upon investigations of many scholars (see his bibliography, 381–86) summarizes the prevailing view with regard to the historical background of American Spanish in these words:

Zamora adds the observation that in colonial times this sustained influence was exerted, or mediated, by the viceroyal courts. Literary Spanish, highly standardized by 1600, continued to have its effects upon the language of the several countries after they achieved independence.

In a general way, American Spanish and American English stand in the same relation to their European sources. Some dialectal differences imported at the time of the settlement survive regionally, fashionable innovations in the language of the mother country continued to be imported in certain centers during the Colonial period, and the literary language of the European homeland is an ever present dominant influence.