10.

Parallel Development or Diffusion?

10.1 Introduction

In diachronie investigations of genetically related languages, the relationships established by the comparative method between the several members of the family prepare the ground for an interpretation of these relationships in sociocultural terms.

Innovations shared by some, but not by all, of the languages belonging to the group seem to be the most significant linguistic guide posts to a historical interpretation. The relative chronology of shared changes should be established for each language, if at all possible; and the geographic domains occupied by languages sharing certain innovations must be taken into consideration in preparation for a decision whether a shared innovation developed independently in the languages that have it in common, or whether it was indigenous to one of them and spread to the other(s). Geographic proximity might speak for diffusion rather than parallel development; remoteness in space might favor the assumption of like indigenous development emanating from trends inherent in the shared system. Two examples may serve to illustrate the relevance of areal data in dealing with problems of this kind.

10.2 The Germanic Language Family

The grouping of the Germanic Languages on the basis of regionally restricted phonological and morphological features has been under discussion for a century and a half, beginning with J. Grimm [1819-37] and R. Rask [1818]. A brief history of divergent views, a listing of relevant features with comments, and a useful bibliography has recently been published by W. P. Lehmann [1965], which the reader will want to consult along with the several comparative grammars of the Germanic stock and the histories of individual languages.

The grouping of the Germanic languages can be exhibited effectively by focusing the attention on structurally and/or statistically important innovations shared by two or more, but not all, of the languages of this family. For the sake of simplicity, only the five languages for which sufficient written evidence is available between ca. 350 and ca. 900 A.D. have been included in the scheme: Gothic, Old Norse, Old English, Old Saxon, and Old High German. This simplification will, I believe, not seriously distort the actual situation.

Shared Innovations in the Germanic Languages

1. Generalization of the inflection ־t (of the PIE perfect) in the pret. sing. 2 of ablaut verbs: Go nam-t “took,” ON halp-t “helped.” Cp. 8 below.

2. Formation of intransitive verbs in -na to match transitive verbs in -ja: Go full-na- “become full” ≠ full-ja- “make full, fill.”

3. Replacement of the stem suffix *־jō by -in in the feminine of the present participle: Go nimand-ein “taking.”

4. Development of long plosives in clusters of laryngeal plus *j or *w; Go twaddje, ON tveggia “two” (gen. pl.); Go triggwa, ON tryggva “troth.”

5. Elimination of reduplication in ablaut verbs of class VII, either by complicated phonological developments of old reduplicated forms of the perfect or by descendants of old aorist forms: OE hēt, OHG hiez (Go haihait) “bid, ordered”; OE hlēop, OHG liof (Go haihlaup) “ran.” Relics of reduplication are found in ON and OE. See Hirt 1932: 140–59; Prokosch 1939: 176–82.

6. Lowering of PGc (< PIE ē) to ä, except in Anglo-Frisian: ON sāð, OHG sāt “seed, seeding.”

7. Allophonic palatalization of back vowels before *j, i, ī, which after the loss of these position markers yielded contrastive vowel phonemes: OE fyllan (< *full-ja) “fill,” settan (< *satja-) “set,” fēt (< *fōt-i) “feet.” See Twaddele 1957 < 1938: 85–87.

8. Generalization of the inflection -i < PGc -iz (of the PIE aorist) in the pret. sing. 2 of all ablaut verbs: OHG nām-i “took,” OE hulp-e “helped.” Cp. 1 above.

9. The assimilation of *j to a preceding consonant: OE biddan “beg, ask,” hliehhan “laugh,” fiellan “fell.”

10. Merging of PGc z with PGc r: OE māra “more” (Go maiza), faran “fare” (Go faran). Though merged in Olcel, Runic Norse preserves the contrast until ca. 700 A.D.

11. Loss of final *-r (< PGc -z) in unstressed syllables: OE wulf “wolf” (ON ulf-r, Go wulf-s); OE hulp-e “(you) helped” (see 8 above).

12. Generalization of voiceless -ƿ in the inflection of the pres. pl. 3: OE ber-aƿ “bear” (Go -and, OHG -ant).

13. Spreading of the inflection of the pres. pi. 3 to the first and second person: OE sing-aƿ (we, you, they) “sing”; OS hori-að “hear.”

14. Loss of final *-r (< PGc -z) in personal pronouns: OE wē “we” (OHG wir), OE mē “me,” dat. (OHG mir).

15. Loss of contrastive dative ≠ accusative forms in the personal pronoun: OE mē, ƿē, ūs, ēow; OS mi ~ me, thī, ūs, eu. However, Old Anglian and OS preserve distinct acc. forms ending in -k to some extent.

16. Shifting of the PGc voiceless plosives p, t, k to fricatives in medial position and to affricates initially: OHG offan “open,” wazzar “water,” mahhōn “make”; pfad “path,” zehan “ten,” chalp “calf.”

The regional innovations presented above are of different dates and of varying importance.

Included are innovations shared by two or more Germanic languages as attested by early texts or inscriptions: Gothic 350–600 A.D.; Runic Norse 350–800; Old English, Old Saxon, Old High German 800–1000; Old Icelandic 1100–1200. The appearance of these shared innovations cannot be dated absolutely. Some may have occurred in the first centuries of the Christian era, others as late as 800–900. Their relative chronology can be inferred with some probability from their regional dissemination.

Reading the table presented above from top to bottom, we travel from the Baltic in a southwesterly direction through central and western Germany to the Alps. On the shores of the Baltic, Gothic and Norse developed some innovations before the Goths migrated southeastward to the Black Sea (features 1–4). In central Germany, along the North Sea, and in southern Germany other innovations are shared by English, Old Saxon, and High German (features 8–11). Some West Germanic innovations (features 5–7) appear also in Old Norse, which shares other innovations with Gothic. In the northern section of the West Germanic area, Old Saxon, Old English, and Frisian jointly introduced some innovations (features 12–15). In the south, a far-reaching shifting of consonants (feature 16) is shared by all of the High German dialects.

On the basis of shared regional innovations one can construct this family tree:

This plan neglects the features shared by North Germanic with West Germanic. However, such a simplified scheme, taken as such, has its useful function in orienting the discussion of the history of a family of languages. It must not be taken too literally.

The regional innovations presented above are of very uneven value from the structural and/or the statistical point of view. Among the most important ones are features 1, 5, 8, and 13, which involve the morphology of large classes of verbs. These can be taken as convenient diagnostic features of the several language groups, each of which is characterized by three or four of the regionally restricted innovations displayed in the table. The fact that their dissemination is coextensive with other regionally restricted innovations makes them representative of the several language or dialect groups.

The regional grouping of innovations poses diachronie problems. For example: (1) What is the relative chronology of innovations 8 to 11 shared by all West Germanic languages? (2) Did these features develop independently in prehistoric stages of Old English, Old Saxon, and Old High German, or did they originate in one of these dialects and spread to the others?

In dealing with innovations of recent times, dated localized documents can furnish evidence for establishing the chronology; knowledge of the precise areal spread of an innovation as established by selective sampling can help us decide whether a shared innovation developed independently in the several dialects or was diffused from dialect to dialect. Finally, and most importantly, the location of a dominant sociocultural center may enable us to determine with some probability the focal area in which the innovation originated and from which it spread. When evidence of these types is lacking, no probable answer can be given to the questions raised above. This is one of the important findings of modern area linguistics, which should discourage hasty decisions in dealing with shared prehistoric innovations.

In passing, it is of interest to note that some shared regional retentions of Proto-Germanic features fall into the same areal patterns as shared innovations. Thus, the sequence -ngw- is preserved in Gothic and Norse: Go siggwan, OIcel syngva “sing”; OE gān “go,” dōn “do” have corresponding forms only in OS and OHG; and OE he, hēo, hit “he, she, it” have counterparts only in Old Frisian and Old Saxon.

10.3 The Indo-European Language Family

Structurally important innovations shared by two or more languages of the Indo-European family form the chief basis for establishing degrees of kinship between them and tracing their diachronie filiation. The evidence for such joint innovations, gathered from texts far apart in time, has been accumulated by scholars for a century and a half and was conveniently summarized by K. Brugmann [1904: 2-27] and by A. Meillet [1907] at the turn of the century.

Chief among shared innovations are the following:

1. Sibilation of the PIE palatals // in Indo-Iranian and Balto-Slavic (the so-called satəm languages), and merging of the PIE labio-velars /k^w, g^w, g^wh / with the pure velars / k, g, gh /.

2. Merging of the PIE palatals with the pure velars and the preservation of contrastive labio-velars in Greek, Italic, Celtic, and Germanic (the so-called centum languages).

3. Merging of the aspirated voiced stops / bh, dh, gh / of PIE with the simple voiced stops / b, d, g / in Iranian and Balto-Slavic; also in Celtic.

4. Development of the “augment” as a verbal prefix in the formation of a past tense in Indo-Iranian and in Greek.

5. Unvoicing of the PIE aspirated stops / bh, dh, gh / to affricated stops in Greek (written φ θ, χ) and to fricatives in Italics (written j:, h).

6. Merging of PEE perfect and aorist forms in a past tense form in Italic, Celtic, and Germanic, thus eliminating the category of aspect.

A striking aspect of these innovations is their areal patterning, which is shown in the accompanying table. Feature 1 is shared by all of the eastern, feature 2 by all of the western languages. The other features are common to two or three contiguous language areas: 3 to Balto-Slavic and Iranian, 4 to Indo-Iranian and Greek, 5 to Greek and Italic, 6 to Italic, Celtic, and Germanic.

This areal patterning of regionally shared innovations furnishes the most important leads for the consideration of the relationships between the Indo-European dialects and their filiation in prehistoric times. Did proto-Indo-European first split into an eastern and a western branch, as is traditionally assumed by inference from the areal behavior of features 1 and 2? Did innovation 3 develop independently in Balto-Slavic and in Iranian, or did it spread from one to the other through intercommunications? Did the augment (feature 4) arise independently in Indo-Iranian and in Greek after their geographic separation or while they were still in contact? Is the formation of a past tense out of earlier perfect and aorist forms in Italic, Celtic and Germanic evidence for an early common development or does it arise independently in the three languages?

Such are the problems raised by the areal evidence. To answer them is not so simple. Regionally shared retensions of PEE features must be taken into consideration to propose probable solutions and the behavior of the several languages during their historical period may shed some light on earlier processes.