Any casual observer of the Indo-European languages will have noted the striking similarity that exists among the various words for the nationality German, for example the following:

Russ.		nemets
Span.		alemán
Lat.-A		germanus
Lat.-B		teutonus
Ital.		tedesco
Ger.		Deutsch1

The following similarities are immediately evident:

(1)		All begin with an alveolodental consonant (ignoring, for the moment, the prefixes of Latin-A and Spanish).
(2)		With the single exception of Spanish all show some kind of sibilant as their final consonant or in their final consonant cluster.
(3)		With the exception of German and, superficially, of Latin-A, all contain the general phonotactic structure C1V1C2V2C3.

In fact the only major schism to be found among the forms is that between the sonorant-dominated forms (Russian, Latin-A, Spanish) and the obstruent-dominated forms (Latin-B, Italian, German). The significance of this dichotomy will be explained below.

Strangely enough, among the reams of work done to date on the reconstruction of Proto-Indo-European, there has been a singular neglect of the lexeme German—I say strangely because semantically the word refers to a major linguistic branch of the Indo-European family itself!

Without wasting any more time on preliminaries, I propose to get right to the matter of reconstructing the proto-form of the word and to the sound changes that have transformed it into the words we find in the above-mentioned modern and classical languages. In the process, I will show just how much the comparative method is capable of, and also resolve once and for all some phonological controversies that have heretofore gone a-begging for historical evidence to confirm their psychological reality.

The similarities cited above can be better appreciated by being displayed in a form such as Table 1.

As I already pointed out, Russian, Spanish, and Latin-A favor sonorant consonants in the root structure, while German, Italian, and Latin-B favor obstruents. Since this division also represents geographically a peripheral/central split, we may consider the sonorant (peripheral) forms to be remnants of the original form, while the obstruent (central) forms suggest a later innovation.

Among the sonorant forms, all have the medial m, and two out of three show the sequence man. As for the initial consonant, it was obviously an alveolar sonorant, namely, r, l, or n. Since it is unlikely that a word that already contained two nasals would assimilate another sound for a third nasal (on the contrary, as I show below), it seems clear that the initial nasal of Russian represents the original form: n-V-man.

Although three of the forms have e for V₁, again on the basis of probably natural sound changes—as it is more likely that the difficult diphthong eu was simplified to e than that e was complicated—I favor an original diphthong eu. The spelling of Ger. Deutsch independently confirms this choice.

In the matter of suffixes, the pervasiveness of the final sibilant in most forms strongly suggests that the -us suffix of the two Latin forms was present in the proto-form.

The proto-form was, then, undoubtedly *neumanus.²

There remains now only to follow the vicissitudes of *neumanus into the daughter languages. Let us begin with the period of common development. Most of the changes of the early period are self-evident in Table 2.

A few comments may help to guide the reader through the table.

Prefixing. In two of the languages, Latin-A and Spanish respectively, the word undergoes a form of folk morphosemantic assimilation or dissimilation, respectively, which we shall call, respectively, homeopathic and allopathic prefixing. This means simply that in the first case a word referring to a language or ethnic group takes on a prefix commonly associated with that group—in this case, the Germanic prefix ge-; and conversely, in the second case (remembering that Spain faced both Germanic-speaking and Arabic-speaking invaders), the word for German took on a prefix from the geographically opposite superstratum, namely the Arabic article al-.

Vowel deletion. For most speakers, the diphthong eu was, frankly, difficult to pronounce. While Latin-A deleted it altogether, other languages deleted only the u; this latter reaction, however, led to the overreaction of deleting all u in the word, diphthongal and innocent alike. German had a larger problem on its hands: a triphthong; this was resolved by deleting a, but here also overreaction carried vowel-deletion through the remainder of the word.

Nasal dissimilation. The dissimilation in Latin-A was perfectly predictable (cf. can-men > carmen). To understand the process in Latin-B and Italian, however, we must bear in mind that these languages abhorred the triple nasal as Nature abhors a vacuum. In addition, both languages had changed little before this point and had to make up for lost time. The solution was radical, by present-day standards. Dissimilation applied to all but the last nasal of the word, and it affected not only nasality but also voicing. The sound change must be formalized as follows:³

+nas +voice

→

-nas -voice

/  ___  (X)  N  (X)  #

Ns-resolution. The cluster -ns was simplified in Spanish by deleting s; other languages epenthesized a homorganic stop, a process not unfamiliar to the punning photographess who sang “Someday My Prints Will Come”.

After the period of common development, the processes in German German that altered the diphthong’s pronunciation but left its spelling alone and that palatalized the clustered sibilant are well known. Likewise in Latin-B, it takes little imagination to discern the process of progressive labialization whereby teupanus > teuponus, nor the bilateral assimilation of p by the flanking dentals, for teutonus (Q.E.D.). Of more interest is the sporadic metathesis of Ital. tepets > teteps, and the secondary metathesis that gave tetesp (this successive metathesis involving three elements, by the way, has been dubbed shellgame metathesis, for obvious reasons). Even with these advances, however, the word no doubt had a “foreign” sound to the average Italian speaker, and it is little wonder that the word was now “naturalized” by intervocalic voicing and by addition of the typically Italian suffix -o. It is the final development of Ital. *tedespo to the modern tedesco that has the greatest implications for phonological theory, however, as it involves a dissimilation of the feature [anterior], thus settling definitively the question of this much-debated feature’s psychological reality. The rule, simply stated, is as follows:

[+ant]  →  [-ant]  /  [+ant]  (X)  [+ant]  (X)  __

In conclusion, I wish to point out that the real challenge in *neumanus and its congeners lies not in the Indo-European family, where relationships are fairly well established, but in other language families not generally recognized as related. The following list merely scratches the surface of the possibilities:

Hungarian		nemet
Tagalog		Alemán
Thai		yermun
Turkish		Alman
Zulu		Jalimane

If the possibilities of using the lexeme German to establish language relationships has been merely neglected within the Indo-European family, outside that family it would seem that today’s linguists have outright actively avoided it!

Footnotes

	My Fair Linguist, or Male-Pygion—Tom Ernst and Evan Smith
	Speech Errors as Evidence for Historical Generative Phonology—Joseph Paul Stemberger
	Son of Lingua Pranca Contents