Wiktionary:About Proto-Germanic

Proto-Germanic is the ancestor of all Germanic languages, old as well as modern. It is a descendant of Proto-Indo-European. It was spoken in north mainland Europe and southern Scandinavia, more or less during the time of the Roman Republic and also in dialectal form during the early period of the Roman Empire (up till about the 1st century CE). No written records of Proto-Germanic exist except for the disputed word harja (ᚺᚨᚱᛃᚨ) in the Vimose inscription from 160 AD (likely too late to be considered Proto-Germanic, and considered here to be Proto-Norse), but the words and grammar of the language have been reconstructed by linguists.

Entries, categories and templates
Being an unattested language, Proto-Germanic is not included in the main namespace. The name of every entry begins with "Reconstruction:Proto-Germanic/" followed by the name of the entry itself. Templates that need the page name can't use to retrieve the name of an entry, but they can use  instead.

Although unattested, Proto-Germanic was a language like any other and is treated as such on Wiktionary for the most part. Proto-Germanic has its own code on Wiktionary for use in templates, which is. This code can be used in most templates. When linking to Proto-Germanic terms, the word should be prefixed with  to indicate that the form is reconstructed. This also tells templates to link to the reconstruction instead of the main namespace.

Proto-Germanic entries are full entries and can have the same sections as other languages, including etymology, derived and related terms, conjugation, declension, and so on. Etymology sections are formatted as usual, and can link to earlier forms of a word in a different (proto-)language, such as PIE, in which case you can use  to categorize correctly. Proto-Germanic has its own categories for parts of speech, etymologies, morphemes etc. and the main category is Category:Proto-Germanic language.

One point where Proto-Germanic entries differ from regular entries is that they must always have a "Descendants" section. This is required because they form the basis for reconstruction of Proto-Germanic entries, and without them there would be no way to judge the validity of the reconstruction. They can be considered a form of "poor person's reference", although of course proper references are no less important.

If a form is not attested in a language, but it can be reconstructed based on a form appearing in a descendant of that language (for example when Old English is unattested, but Middle English is attested), then its reconstructed form(s) may be provided with an asterisk * as is usual for unattested terms. If a certain descendant does not exist, do not reconstruct it if it does not itself have a descendant: for example, don't reconstruct Old English if no Middle English or Modern English descendant exists. Most of the time, Gothic should not be reconstructed at all as it has no direct descendants, although borrowings (if the Gothic donor term can be ascertained) may serve as a rare basis for reconstruction. As a special exception, it's preferred not to include Frankish reconstructions unless a language other than Dutch has a term derived from it (for example French or Medieval Latin).

When descended in the usual, linear fashion, descendants are always listed in the following order, just for the sake of consistency.


 * Proto-West Germanic
 * Old English
 * Middle English
 * English
 * Scots
 * Yola
 * Old Frisian
 * North Frisian
 * Saterland Frisian
 * West Frisian
 * Old Saxon
 * Middle Low German
 * Low German
 * Dutch Low Saxon
 * German Low German
 * Plautdietsch
 * Old Dutch
 * Middle Dutch
 * Dutch
 * Afrikaans
 * Limburgish
 * West Flemish
 * Zealandic
 * Old High German
 * Middle High German
 * German
 * Rhine Franconian
 * Pennsylvania German
 * Volga German
 * Central Franconian
 * Hunsrik
 * Transylvanian Saxon
 * East Central German (Upper Saxon, Silesian German, etc.)
 * East Franconian
 * Bavarian
 * Cimbrian
 * Gottscheerish
 * Hutterisch
 * Mòcheno
 * Alemannic German
 * Swabian
 * Sathmar Swabian
 * Colonia Tovar German
 * Luxembourgish
 * Vilamovian
 * Yiddish
 * Zipser German


 * Proto-Norse
 * Old Norse
 * Icelandic
 * Faroese
 * Norn
 * Norwegian
 * Bokmål
 * Nynorsk
 * Jamtish
 * Elfdalian
 * Old Swedish
 * Swedish
 * Old Scanian
 * Scanian
 * Old Danish
 * Danish
 * Old Gutnish
 * Gutnish
 * Gothic
 * Crimean Gothic
 * Vandalic

If a term derives non-linearly, it should be sorted below its etymon. In this example, taken from, the English word shtetl derives from Yiddish, not from Old English:

Phonology, spelling and notation
Proto-Germanic on Wiktionary is represented as follows; the transcription in IPA is shown under each symbol.

Consonants

 * Voiced fricatives and voiced plosives were allophones of each other. For this reason we should not make a distinction between them in spelling, as many sources and references do not either. This means that we do not use the letters ƀ, đ, ð or ǥ, nor bh, dh or gh, but instead b, d, g.
 * The voiceless dental fricative [θ] is represented with the letter thorn þ. This follows established practice in Germanic linguistics.
 * [h], [x] and [χ] were all the same phoneme. As all early attested Germanic languages and many sources on Proto-Germanic use the letter h to represent the phoneme, we should do so here, irrespective of the actual phonetic realisation.
 * The labialised velars are represented by a following letter w: hw, kw, gw. The alternative possibility, a superscripted ʷ, may be more accurate, but Germanic did not contrast labiovelars with velar+w sequences, so there is no confusion possible. And a regular w is easier to type.

Vowels

 * Long vowels are represented by macrons over the letters: ā, ē, ī, ō, ū.
 * Nasal vowels are represented by an ogonek (ą, į, ų, ǭ etc.). Other sources may use the tilde (ã, ĩ, ũ) or the letter n following the vowel (an, in, un), but the ogonek is probably preferred for a few reasons. The tilde tends to look like a macron, making them hard to distinguish at low resolutions. The final n meanwhile, while immediately clear, can be confused with a proper word-final -n, which occurs in some places where a final -t has elided (notably in the 3rd person plural endings of verbs and in some numerals).
 * The long nasal vowel that developed from a vowel plus the sequence nh (through compensatory lengthening) is represented as simply the short vowel (a, i, u) followed by nh. That is, the nasality is not indicated, but is implied by context.
 * The phoneme named ē1 in various sources should be represented as simply ē, as in Gothic and several sources. The letter æ is not used, nor is the combination ae.
 * The status of ē2 is open to discussion, but whenever it can be reconstructed it is represented as ē₂ (with a small 2 character, it can be found on the Miscellaneous tab at the bottom of the edit window).
 * Overlong/trimoraic vowels (as posited by D. Ringe and others) are represented with a circumflex (ê, ô). This seems like the most workable solution. Ringe uses one macron over another in his book, but that is a very unusual character combination which is not widely supported, not even in Unicode.

Diphthongs

 * Diphthongs in Proto-Germanic are combinations of a vowel followed by one of the semivowels -j- or -w-.
 * The semivowel is written -i- or -u- when followed by a consonant.
 * ...except when another semivowel of the same kind immediately follows: *ajją, *triwwiz, *brewwaną.
 * -w- remains when -j- follows (*tawjaną) but -j- becomes -i- when -w- follows (*saiwiz).
 * The first vowel can be any vowel except u: ai, au, eu, iu.
 * There are also diphthongs with a long vowel, but these are rare: ēi, ēu, ōi, ōu.
 * ēi probably merged with the close-mid long vowel ē2.
 * The diphthong -ei- does not exist, it is written -ī- (see below).
 * A vowel followed by a non-high vowel is not a diphthong, but rather a sequence of two vowels, and consists of two syllables (*fraetaną 4 syllables, *wēaną 3 syllables). Such combinations often contracted into long vowels or dipthongs later though, especially when the first vowel was unstressed.
 * However, when the first vowel is a long high vowel, this can also be seen as either a high vowel+semivowel glide. Going by common practice, the glide is used for -ī- (frijōndz) but the long vowel is used for -ū- (būaną). In theory these can also be seen the other way (frīōndz, buwaną), but this is not done here.

Stress
Stress should not be marked at all, because it's not phonemic and always occurs on the root syllable. In compounds, stress is placed on the first part of the compound. Nouns and adjectives with adverbial prefixes are considered compounds, but verbs are not, just as in the modern Germanic languages.

Which Proto-Germanic?
Proto-Germanic was not a static language, it underwent many changes throughout its history. To clearly represent the language however, and to avoid confusing forms that only differ in time, it would be a good idea to pick a particular stage and stick with it. Here are the most important changes generally accepted to have occurred by late Proto-Germanic, in a somewhat chronological order. For a verifiable source, From Proto-Indo-European to Proto-Germanic by Don Ringe can be used. See also Proto-Germanic and Phonological history of English.

Pre-Germanic refers to a stage of Germanic before the loss of word-final non-high short vowels and Grimm's law, essentially a Proto-Indo-European centum dialect that had lost the laryngeals (between consonants, they were vocalized into *ə), coloring them in the process (as in most Indo-European languages) with compensatory lengthening if they occurred after vowels, and created sonorant diphthongs with an epenthetic *u inserted before PIE syllabic resonants.


 * Unstressed word-final -a and -e disappear.
 * Grimm's and Verner's law. These two shifts are generally held to define Germanic as such.
 * Germanic spirant law: bt/pt > ft, dt/tt/þt > ss, gt/kt > ht. In verb paradigms ss was replaced by st by analogy with other verbs, and similarly for some suffixes.
 * o and a merge as a, ō and ā merge as ō. (New ā formed from contraction of older -aja-, see below)
 * Word-final -m > -n (including stressed syllables).
 * In unstressed syllables, final -n disappears and nasalises the previous vowel. This precedes the following change, so that in an -nt sequence, -n remains.
 * Unstressed word-final -t disappears.
 * Unstressed e changes to i except before r. This precedes the following change, because the newly-formed i's were later able to cause mutation themselves.
 * I-mutation of e to i before following i or j. This also turns a preceding eu into iu.
 * ei > ī as part of the previous change. While it may be possible to find verified sources with ei, the problem is that we can't find sources for all historical instances of ei. To avoid giving the impression that there is a meaningful distinction between ei and ī, we should use ī for all cases.
 * Sievers' law: -j- and -ij- alternate depending on the weight of the preceding syllable(s). -j- occurs after light syllables, which have a short vowel and end in a single consonant. -ij- occurs after heavy syllables, which have a long vowel, diphthong or end in two or more consonants.
 * e > i before nasal + consonant.
 * Loss of intervocalic j. iji > ī, ji > i, ijV remains. All other VjV > VV (the two vowels either contract into a long vowel, or form a diphthong). In particular, aja > ā (a rare phoneme, but occurs in class 3 weak verb endings and in the short form of 'stand'). Any remaining instances of j are either word-initial or preceded by a consonant or i.
 * Note that j survives after diphthongs ending in -u (such as Gothic taujan, niujis), as diphthongs were still structurally vowel+semivowel combinations. The phonemic structure was -wj-, not -uj-, and this is demonstrated by the outcome of the West Germanic gemination, which also geminated -wj- to -wwj-.

Changes we should ignore:


 * A-mutation of u to o and i to e before following a. The extent to which this change occurred differs between dialects (Gothic, for one, didn't have it at all), and is mostly incomplete even where it does occur. This means that our representation of Proto-Germanic lacks short o as a phoneme.
 * Loss of final -i and -u. There is no secure dating on this change and it may well have occurred after the languages already split up (West Germanic, for one, preserves both in certain contexts).
 * Loss of n before h with lengthening and nasalisation of the preceding vowel. This change was a very late one in Germanic and was most likely not yet phonemicised.

Grammar points

 * The infinitive ending of class 2 weak verbs is -ōną, not -ōjaną. The latter ending is only attested in the northern West Germanic languages, and in any case the -j- would not have survived the loss between vowels. Most likely the -j- was an innovation based on the pattern of class 1 weak verbs, and could have progressed something like this:
 * Original situation: -jan/-ida, -ōn/-ōda
 * -i- is lost in class 1 verbs: -jan/-da, -ōn/-ōda
 * -j- is introduced based on the pattern -Vjan/-Vda: -jan/-da, -ōjan/-ōda.

Dialectal forms
Some Proto-Germanic words are only reconstructable for a certain branch of the language. Proto-West Germanic and Proto-Norse are treated as their own language, with the codes  and   respectively. Words that are only reconstructable in one of these branches should not automatically receive a Proto-Germanic page. On the other hand, Proto-Northwest Germanic, the common ancestor of the two, is treated as a dialectal form of Proto-Germanic itself. It does not differ significantly enough in form to warrant treating it as its own language. Terms that are innovations only in Northwest Germanic can be added as Proto-Germanic pages, labelled with.

It should be kept in mind that even if a Proto-Germanic word only has descendants in one branch, this does not mean that the word can't be reconstructed for Proto-Germanic proper. The word may simply have been lost in the other branches. Strong verbs, for example, are almost always inherited words even if they don't survive in all three branches. This is similar to the differences in UK and US English, since a word that is now considered "American" could have simply been lost in UK English, but have still been present in their common ancestor, Early Modern English (for example, in the sense of ‘autumn’). For the purposes of considering what is truly a dialectal word not present in Proto-Germanic, we consider only innovative formations. This means generally words that have been coined through a derivational process that was still productive in Proto-Germanic. Words that were formed by non-productive processes obviously can't be new formations, and must have been inherited.