Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

A Definition of Proto-Germanic: A Study in the Chronological Delimitation of

Languages
Author(s): W. P. Lehmann
Source: Language , Jan. - Mar., 1961, Vol. 37, No. 1 (Jan. - Mar., 1961), pp. 67-74
Published by: Linguistic Society of America

Stable URL: https://www.jstor.org/stable/411250

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/411250?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

is collaborating with JSTOR to digitize, preserve and extend access to


Language

This content downloaded from


180.176.178.34 on Sun, 23 Jan 2022 14:01:57 UTC
All use subject to https://about.jstor.org/terms
A DEFINITION OF PROTO-GERMANIC
A STUDY IN THE CHRONOLOGICAL DELIMITATION OF LANGUAGES

W. P. LEHMANN

University of Texas

It may seem obvious that when we set out to write a grammar of a language,
the definition of that language is one of our first tasks. Further, that since we are
dealing with language, the criteria used in our definition should be linguistic.
These criteria should delimit the language geographically, that is, from other lan-
guages, and chronologically, that is, from earlier and later stages of the same
language. Yet in spite of these minimum requirements for an adequate definition,
our handbooks are not precise in their definitions or explicit in regard to criteria
used for the definitions they give.
Though we are here concerned with the definition of Proto-Germanic, the
problem applies to all languages. Handbooks often delimit a language through th
use of social or political changes rather than linguistic changes. Middle High
German, for example, has been defined as that stage of German which was spoken
between the period of the great medieval German poets and the Reformation. The
Norman Conquest and the introduction of printing have been used to define
Middle English. While linguistic changes may coincide with other social changes
the coincidences are interesting correlations rather than defining features for the
language. We can determine them only after we have selected those differences
in language structure which distinguish successive periods of any language, such
as Middle English in contrast with Old English and Modern English, or different
languages, such as Italian in contrast with French and Sardinian or Low German
in contrast with High German and English.
Imprecise definitions are objectionable in part because they permit unneces-
sary disputes to arise regarding the description of a language. Two inadequate
definitions from grammars in general use may indicate the need for improved
definitions.l Prokosch, CGG 26, contents himself with indicating the location of
the Germanic speakers and adding:2 'Shortly before the beginning of our era, the

1 Since the bibliographical material on Proto-Germanic is easily accessible and well


known, I refer only to a few of the standard texts; further bibliography is listed in them.
E. Prokosch, A comparative Germanic grammar [CGG] (Philadelphia, 1939) lists bibliography
300-302; F. Mosse, Manuel de la langue gotique2 [MIG] (Paris, 1956) lists a small set of works
important especially for Gothic studies 15-17. F. van Coetsem, Das System der starken
Verba und die Periodisierung im alteren Germanischen (Amsterdam, 1956) lists bibliography
83-86. Recent bibliography is available in the pertinent sections of the Bibliographie lin-
guistique and the annual bibliographies of the Modern Language Association.
2 For 'primitive', it is now the convention to use Proto-, to avoid the connotation carried
by the nontechnical use of 'primitive'. The prefix pre- is used to indicate a prior stage to
any given language when the prior stage is not capable of more precise definition, or when
more precise definition is not pertinent for the moment. As in the quotation from Mosse
the term 'common' (Germanic) has often been used in dealing with features found in all
dialects which have developed from a given language but which may not have a distinct
67

This content downloaded from


180.176.178.34 on Sun, 23 Jan 2022 14:01:57 UTC
All use subject to https://about.jstor.org/terms
68 LANGUAGE, VOLUME 37, NUMBER 1 (1961)

Germanic group appears to have been a fairly homogeneous linguistic and cul-
tural unit. This is the period that is termed Urgermanisch, Primitive Germanic.'
Mosse, also a clear and careful writer, makes a similar statement, MIG 19: 'Les
langues germaniques sont les formes diverses prises, au cours de l'6volution his-
torique et du morcellement g6ographique, par une langue commune parl6e
jusque vers le ddbut de l'ere chr6tienne dans le Sud de la p6ninsule scandinave,
le Danemark et la plaine de l'Allemagne du Nord et que l'on appelle le proto-
germanique (all. Urgermanisch) reservant le terme de germanique commun (all.
Gemeingermanisch) pour d6signer les faits post6rieurs communs A toutes les an-
ciennes langues germaniques.' Neither Prokosch's definition nor Mosse's meets
the requirements I have set, for each is little more than an impressionistic sugges-
tion. Yet other handbooks have failed to give better definitions. After several
decades of Germanic study it seems desirable to define Proto-Germanic pre-
cisely. Upon proposing such a definition I will note some disputed problems in
Proto-Germanic grammar and suggest how my definition may aid in clarifying
them.
For linguistic purposes, establishing the relative chronology of Proto-Germanic
is more important than determining the location of its speakers. For both aims
our only usable evidence to the present is linguistic; until we find inscriptions
which enable us to relate prehistoric cultures of northern Europe with prehistoric
linguistic communities, attempts to locate the speakers of pre-Germanic or post-
Indo-European dialects in northern Europe are completely speculative, if in-
triguing. Only on the basis of the available linguistic evidence can we make
reliable statements about the location and the dates of the community of Proto-
Germanic speakers.
In framing our chronological definition of Proto-Germanic we use a technique
similar to that which is used to distinguish Germanic from other Indo-European
dialects.3 Rather than choose natural boundaries-rivers, seas, mountains-to
bound Proto-Germanic and separate it from neighboring dialects such as Baltic,
Celtic, Greek, Italic, Slavic, we choose linguistic data, of which the following
are samples which can easily be supplemented.
Phonetically we can distinguish Germanic from Baltic, Celtic, Italic, Greek,
and Slavic by the presence of [f 1' x], which correspond to [p t k] in those dialects

etymon in that language. Handbooks label umlaut Common Germanic but not necessarily
Proto-Germanic. Since the label is confusing in suggesting a further distinct phase of a
language, I avoid it entirely; if necessary we may describe a feature as common to various
segments of a linguistic group, such as Germanic.
3 The terms geography and geographical in their linguistic usage are to be taken as in-
dicating relative distribution of a segment of speech. In spite of possible misunderstandings
I maintain these terms with other terminology based on 'the study of the earth' because
they are established. Yet terminology based on language distribution or linguistic dis-
tinctiveness would be less likely to mislead than that based on the concrete classifications
of a physical science. An example is the persistence of the classification into North, East,
and West Germanic, and the positing of proto-forms with these labels in spite of the absence
of proof for such subgroups subsequent to Proto-Germanic. The attractiveness of a favored
tripartite grouping based on the points of the compass seems to outweigh more complex
classification based on available data.

This content downloaded from


180.176.178.34 on Sun, 23 Jan 2022 14:01:57 UTC
All use subject to https://about.jstor.org/terms
A DEFINITION OF PROTO-GERMANIC 69

and in Proto-Indo-European (or to clearly definable structural points


environments, such as zero for PIE p in the Celtic dialects).
Phonemically we can distinguish Germanic by the loss of contrast b
reflexes of Indo-European voiceless stops and voiced aspirated stops a
accented vowels, its preservation elsewhere. A pair illustrating the lo
trast is Goth. sibun 'seven' with reflex of PIE p and Goth. giban 'give' wi
of PIE bh; a pair illustrating its maintenance is Goth. faihu 'cattle' w
of PIE p and Goth. bairan 'bear' with reflex of PIE bh. In Celtic, Balt
Slavic the contrast between PIE b and bh was completely lost; in Italic it was
partially maintained; and in Greek it was maintained except when a reflex of
another aspirate occupied a syllable contiguous with that having a reflex of PIE
bh. A complex realignment like that of PIE p: bh, t: dh, k: gh in Germanic
provides excellent means for geographical definition, for even the varied develop-
ments in Italic and Greek fail to coincide with it, not to speak of the simple
changes in Celtic, Baltic, and Slavic.
Morphologically we can distinguish Germanic by the development of two ad-
jective declensions and of a weak preterite marked with d.
Such data for distinguishing Germanic geographically we find in our hand-
books, although since our handbooks dealing with Proto-Germanic are based on
neogrammarian theory they fail to make use of the strongest distinguishing cri-
teria at the phonological level, which are structural. Data like those cited here so
clearly distinguish Germanic from neighboring Indo-European dialects that
questions rarely arise in identifying Germanic material of any extent as distinct
from that in other Indo-European dialects. To ensure a satisfactory description
of Proto-Germanic we need to use similar means for delimiting it chronologically.
In defining a language chronologically, we should also rely on structural differ-
ences, but now the differences we select are carried out within a given territory
and a given linguistic tradition. Again the usefulness of our definition depends on
the significance of our criteria. One measure of their significance is the differen-
tiating effect they have had on the language at the time of their occurrence and
on its subsequent stages. Accordingly we choose criteria which affect more than
one of the systems in a given language.
The criteria I choose for defining Proto-Germanic in its inception and in its
conclusion are changes in its suprasegmental system and their results. The incep-
tion of Proto-Germanic I place at the completion of its shift from free pitch accent
to a stress accent fixed on the initial or stem syllable of words. This accent shift
entailed a major shift in the segmental system, the first major structural change
among the phenomena associated with Grimm's law. For as the accent shift was
completed, the distinction between two of the three orders of Indo-European
stops was broken down medially, under certain accentual conditions, as in Goth.
giban of the Indo-European bh-order and sibun of the p-order.4 Moreover, the
new alignment in the Proto-Indo-European voiceless stop order was exploited
morphologically. Its possibilities may be illustrated by the contrast between the
4 For the traditional terms order, referring to classes determined by manner of articula-
tion, and series, referring to classes determined by place of articulation, see C. D. Buck,
A comparative grammar of Greek and Latin 33 (Chicago, 1933).

This content downloaded from


180.176.178.34 on Sun, 23 Jan 2022 14:01:57 UTC
All use subject to https://about.jstor.org/terms
70 LANGUAGE, VOLUME 37, NUMBER 1 (1961)

medial consonant of the infinitive in OE cweJan, OS quethan, OHG qued


that of the past participle, OE cwedan, OS quedan, OHG quetan;5 these we
available in word formation, e.g. OE deat 'death' versus OE dead 'dead'. Since
the shift in accent entailed structural shifts in the various systems of Proto-
Germanic, it constitutes a major break in the history of Germanic and merits
recognition as the essential dividing mark between pre-Germanic and Proto-
Germanic.
The structural shift I choose to define the conclusion of Proto-Germanic is a
result of the accent shift: the loss of the final short low vowels /e a/ when they
did not have primary stress. Their loss is the only one finally among that of the
Germanic short vocalic segments [i e a u] which is general to all the Germanic
dialects. Runic -gastiR and magu indicate retention of -i and -u to the time of
North Germanic; OE wine vs. giest, sunu vs. feld show reflexes of PGmc. [i u]
preserved even into the dialects.
With the loss of weakly stressed final [e a] the distribution of Germanic vocalic
phonemes was greatly altered. Moreover, morphological categories were modi-
fied, such as the first and third persons preterite of strong verbs, which coalesced
as a result of this phonological change, e.g. Goth. wait 'I know, he knows' in
contrast with Gk. otda 'I know' and oide 'he knows'. Again the selected struc-
tural change affects various systems of the language.
Proto-Germanic I then define as that stage of Germanic which was spoken
between the time of the Germanic accent shift and the loss of /e a/ when final
and weakly stressed. A grammar of Proto-Germanic will be a description of this
stage of Germanic. Even a short sketch of the various systems of Proto-Germanic
would be unduly long. Here I note briefly some disputed questions which my
definition may clarify.
One of these questions is determination of the reflexes of PIE /e/ from position
to position. Throughout the history of Germanic studies there have been sugges-
tions that PIE [e] and [i] coalesced in Germanic.6 We may cite evidence that at
least in some positions they did not.
When weakly stressed and final, [e] did not become [i] in Proto-Germanic. For
if final weakly stressed [e] had become PGmc. [i], it would have remained after
short root syllables, as did the reflex of PIE [i], in Old English, Old Saxon and
Old High German, e.g. in OE men, later mere 'sea', OS OHG meri < PIE [meri],
cf. Lat. mare, gen. pl. marium. If PGmc. -e had become -i, the imperative 2d sg.
of faran in Old English should have been *fsri.

5 More significant morphologically was the coincidence of the past participle marker,
from Proto-Indo-European t, and the preterite tense marker, from Proto-Indo-European
dh, in weak verbs. Only after these fell together can the Germanic verbal system have
developed as a tightly knit subsystem with nominal forms included in the paradigm. But
since some scholars still seem to favor theories which derive the preterite marker from
PIE t rather than dh, ascription of a single preterite and preterite participial marker in
weak verbs to one of the results of the Germanic accent shift would not be universally ac-
cepted and is therefore omitted from the body of the text here.
6 For two discussions of the problem see W. F. Twaddell, The prehistoric Germanic
short syllabics, Lg. 24.139-51 (1948), and J. W. Marchand, Germanic short *z and *e: Two
phonemes or one, Lg. 33.346-54 (1957).

This content downloaded from


180.176.178.34 on Sun, 23 Jan 2022 14:01:57 UTC
All use subject to https://about.jstor.org/terms
A DEFINITION OF PROTO-GERMANIC 71

Moreover, -e- of root syllables did not become -i- before final unst
it should have, if -e had become -i); cf. ONorw. mek, OE mec, and
'me'.
Though this evidence is negative, the absence of any final -i in Old English,
Old Saxon, or Old High German which could be a reflex of pre-Germanic -e is
completely convincing on the loss of final -e before weakly stressed -e became -i
in Germanic. Accordingly by my definition of Proto-Germanic, final -e remained
unchanged and in contrast with final -i throughout the Proto-Germanic period.
Although proper names must be used with caution, they comprise our only
Proto-Germanic material and they can serve to support conclusions based on re-
constructions. The oldest proper names preserve -e- in stressed syllables, even
before vowels which subsequently became -i- and caused change of e to i, e.g.
Tacitus' Venedi as compared with OHG Winida, Erminones as compared with
OHG Irmin-sul, Gepidos as compared with OE Gife&as. Forms which we can re-
construct with assurance, such as PGmc. meke, and names preserved in classical
materials therefore support the conclusion that throughout Proto-Germanic
stressed e was maintained distinct from i.
Yet since the evidence in stressed syllables has been interpreted both for and
against a four-vowel system, the crucial evidence is from unstressed syllables.
There is little question that -i was preserved in the Proto-Germanic etymon of
Goth. man-, OE meri 'sea', etc., which we may reconstruct PGmc. [mari]. On the
other hand, -e must have been maintained in such 2d sg. imperative forms as
PGmc. [gebe], the etymon of Goth. gif, OE gef, gief, etc. PGmc. [i] and [e] then
contrasted, and must be assigned to separate phonemes.
A second disputed question is the allophonic position of PGmc. [i]. Previously
I have stated that, while the Proto-Germanic short vowel phonemes were /e/
and /a/, [i] was a member of the /y/ phoneme, which like the five other Indo-
European resonants was maintained into Proto-Germanic with the threefold
allophonic variation of Proto-Indo-European, Lg. 31.355-66 (1955). Arguments
presented against this view have been weakened by not being made in a proper
chronological frame. We may recall how Streitberg failed to note the relevance
of chronology in discussing the Proto-Germanic e: i situation. He assumed that
the inclusion of Goth. bidjan, OIcel. bi&ia, OE biddan among strong verbs of the
fifth class was an argument for the coalescence of e and i in Proto-Germanic; for
he interpreted its root as a reflex of PIE bhydh- not bhedh-. The inclusion of a
Gmc. bid- in the fifth class, and the formation of a preterite bad seemed to him
an indication that PGmc. i was an allophone of e, Urgermanische Grammatik 53
(Heidelberg, 1896). Yet bad- may have developed subsequently in the various
dialects because of its root structure CVC, as the preterite of the Germanic form
of Lat. scribere was made on the basis of a root structure sCrVyC. Accordingly
the realignment of bid- is of little significance for understanding the status of
Proto-Germanic e: i; it indicates however the intricacies of Germanic vocalic
and morphological development, and the necessity of viewing Proto-Germanic as
a language with the complexities of languages we know in greater detail.
To my earlier suggestion that the Indo-European allophonic variation of reso-
nants, in the patterns described under Sievers' law, survived into Germanic,

This content downloaded from


180.176.178.34 on Sun, 23 Jan 2022 14:01:57 UTC
All use subject to https://about.jstor.org/terms
72 LANGUAGE, VOLUME 37, NUMBER 1 (1961)

Marchand raised a vigorous objection, Lg. 32.285-7 (1956). Dealing with only a
part of the evidence, he contended that Sievers' law was not operative in Proto-
Germanic; he reconstructed etyma of Gothic first class verbs in Proto-Germanic
and Proto-Indo-European, of which I reproduce only the third singular:

*-ye/o- *-eye/o-
LONG STEMS SHORT STEMS LONG STEMS SHORT STEMS

PIE *sagiyeti *kapyeti *gouseyeti *logheyeti


Gmc. *sokiyid/i *hafyid/ki *kausiyid/Ji *lagiyid/Ii
Goth. sokeip -hafjip kauseip lagjib
By my point of view, the post-root vocalism in these forms was rearranged in
accordance with Sievers' law; by Marchand's, it was rearranged by analogy, and
the Sievers'-law variations were no longer in operation.
As an argument that he considers to be in his favor, Marchand cites minimal
contrasts between Proto-Germanic infinitives and participles which were etyma
of Goth. beidan etc. and bidan etc.; he reconstructs the Proto-Germanic forms
*biydanan and *bidanan, labeling the vocalic nucleus of the infinitive stem the
PGmc. reflex of PIE ey. He goes on to assert that if the allophonic situation de-
scribed under Sievers' law had been in operation in Proto-Germanic the two
forms would have fallen together. At the time that Marchand published his
view, van Coetsem stated at some length his opinion that -ey- remained un-
changed until late in Proto-Germanic. But even if he did not accept van Coet-
sem's statement, the difficulty is a figment of Marchand's. The -iy- he posits in
the infinitive stem is his own construction; actually PIE -ey- as syllabic nucleus
coalesced in Germanic with the reflex of PIE z, not with iy. Like the other 'long'
vowels PGmc. z is analyzed as a geminate vowel, -ii-, by Hamp in his article on
Germanic final syllables, Studia linguistica 14.29-48 (1960). The third allophone
of PIE /y/, [iy], like those of the other resonants, never occurred except at syl-
lable boundaries. Accordingly, whether we posit the generally assumed PGmc.
/1/ or Hamp's /ii/, the contrast between the vocalism of infinitive and participle
in Proto-Germanic has nothing to do with the allophonic variation of /y/ in
Proto-Germanic or with the suffix of the first class weak verbs. In the third
singular of weak verbs the forms were PGmc. [kausiyidi] and [lagyidi]; the in-
finitive stem of Goth. beidan in Proto-Germanic was /bid-/ or /biid-/ from
/beyd-/ from PIE bheydh-. Accordingly there is no conflict between the allo-
phonic situation in [kausiyidi] and the vocalic development of PIE bheydh-.
There is further no obstacle to assuming that weakly stressed -e- of the Proto-
Indo-European suffix -eye-, as in *gouseyeti and *logheyeti, had become -i- before
y. Marchand stated that if [e] before [y] had become [i], this change would have
disrupted the allophonic variation of /y/, bringing about a phonemicization of
[y] and [i]. This statement cannot be supported by observation of change in
historical linguistics. As an instance of Marchand's supposed rule we might cite
the allophonic structure of the allophones of American English /d/ initially and
intervocalically after strongly stressed vowels. Intervocalically an allophone of
/t/ is coalescing with an allophone of /d/, as in latter and ladder. Yet the two
phonemes are being kept distinct elsewhere, and no new D phoneme has arisen

This content downloaded from


180.176.178.34 on Sun, 23 Jan 2022 14:01:57 UTC
All use subject to https://about.jstor.org/terms
A DEFINITION OF PROTO-GERMANIC 73

between vowels. For a set of allophones in a phoneme may be e


addition of merging members and still maintain its relationship
of allophones in that phoneme. We may also then assume that
before -y- with no disruption of the /y/ phoneme. Accordingly
of Sievers' law' in short stem verbs like lagjan, the expected iy from
have become [y], giving rise to the forms we find in the dialects
lagjip.
Since the objection which Marchand raised to the Germanic allophonic varia-
tion of /y/ has proved groundless, there is no point in restating here evidence
for positing the retention into Germanic of the Indo-European variation of
resonants in other surroundings. Yet the demonstration, given above, of the
continued contrast in Proto-Germanic between final [e] and [i] (probably even
after PIE -ey- > PGmc. -ii-) supports the assumption that [i] maintained into
Proto-Germanic its allophonic position as a member of the /y/ phoneme. The
two problems discussed are then intricately allied, and must be carefully exam-
ined within a chronological frame.
Defining Proto-Germanic as we have permits a description of it comparable to
those of attested languages in rigor, if not in depth. Its phonological structure I
have sketched elsewhere, JEGP 52.140-52 (1953). In accordance with the defini-
tion given here we must posit for its morphological system full endings, since
final vowels were preserved to the end of the Proto-Germanic period. The dative
plural forms, Vatvims and Aflims, attested in Latin materials-F. Kluge, Urger-
manisch 197 (Strassburg, 1913)-indicate the extent of preservation of nominal
endings. Verbal endings were similarly preserved, as in the first and third singu-
lar of the preterite. Throughout Proto-Germanic, morphological categories were
clearly marked phonologically. Only after Proto-Germanic do we find a falling
together of forms, with greater reliance on syntactic devices to indicate distinc-
tions which earlier were morphologically marked. Our definition therefore assists
us in determining the morphological and syntactic structure of Proto-Germanic,
and in clarifying our description at its various levels, not only the phonological.
Besides assisting us in clarifying problems in any stage of a language, chrono-
logical definitions will also enable us to set up stages of pre-languages rather than
treat them as huge amorphous masses. Proto-Germanic is scarcely to be dated
before 500 B.C., yet a very conservative estimate of Proto-Indo-European would
set for it a final date of at least 2500 B.C. The intervening period of two thousand
years during which some species of Germanic was spoken is generally shrugged
off with the label pre-Germanic, although in the two thousand years after Proto-
Germanic at least three subclassifications are posited for English, German, and
the other Germanic dialects. With structural analyses we can propose subclassi-
fications of prehistoric languages too; such more detailed classification is one of
the most useful and interesting tasks of contemporary historical linguistics. It
will enable us to gain a clearer idea of the development of any given language,
and of language generally, in accordance with the last section of Otto Jespersen's
Language (London, 1922). To deal satisfactorily with the 'development of lan-
guage' in general, to decide whether Jespersen's view of a progression in language
from complex sentence-like entities to individual word-like entities is to be pre-

This content downloaded from


180.176.178.34 on Sun, 23 Jan 2022 14:01:57 UTC
All use subject to https://about.jstor.org/terms
74 LANGUAGE, VOLUME 37, NUMBER 1 (1961)

ferred to Whitney's view of precisely the opposite progression, we will have to


analyze in detail, and with great care, the prehistory of numerous languages.
Three stages in the pre-history of Germanic were proposed by F. van Coetsem
in his capable monograph, Das System der starken Verba und die Periodisierung
im dilteren Germanischen: 1. Urgermanisch, which 'loses itself in Proto-Indo-
European'; 2. Gemeingermanisch; 3. the individual dialects. To the first stage,
characterized by free accent, like that of Proto-Indo-European, he assigned the
changes described by Grimm's law and Verner's law. Yet the strong stress accent
on the first or stem syllable he ascribed only to the second stage. Although further
developments which he places in each period are acceptable, the basic criteria that
he proposes are not, for the sound change described by Verner affected the struc-
ture of Germanic only when the accent was fixed. Moreover, it is something of a
simplification to say that Proto-Germanic loses itself in Proto-Indo-European,
for in the interval between Proto-Indo-European and Proto-Germanic important
changes have taken place at the morphological level. More compact inflections
have been developed, for substantives as well as verbs, but especially for verbs;
for adjectives on the other hand two basically different inflections have been
developed. Further, the basis of inflection has changed, in the verb from an indi-
cation of aspect to tense, in substantives from an indication by case forms of re-
lating a substantive to a given situation, primarily local, to indication of rela-
tionship with other components of the sentence. When these and other changes
are associated with the phonological changes which took place in early Germanic,
we will have a more precise and complex statement of the history of pre-Ger-
manic. The greater complexity will lead us to recognize that the two thousand
years of pre-Germanic were not empty of change; the greater precision will enable
us to provide a more balanced history from Proto-Indo-European to contem-
porary Germanic dialects.
Chronological delimitation of languages, based on linguistic criteria, must be
stated or implicit in all future grammars. The possibility of defining languages
structurally has made obsolete the vague statements given formerly in delimiting
languages such as English or German, dialects such as Bavarian or West Saxon,
or poorly specified stages of either such as Middle English, Old High German, or
Proto-Germanic. Our definitions may be minute or broad. We may choose to
deal with a span of several centuries or several millennia, just as geographically
we may restrict ourselves to a single dialect of English or deal with the whole
range of modern English. Whether the scope of our treatment is small or great,
in the future it must be precisely specified. Such precise techniques have been
evolved for the description of language that we have no further excuse for poorly
supported definitions, including those of prehistoric languages like Proto-
Germanic.

This content downloaded from


180.176.178.34 on Sun, 23 Jan 2022 14:01:57 UTC
All use subject to https://about.jstor.org/terms

You might also like