01
Origins
Proto-Germanic is not written down anywhere. It is reconstructed by the comparative method: line up Gothic, Old Norse, Old English, Old High German and their relatives, work out which differences are regular, and infer the ancestor that explains them all. Proto-Germanic's own ancestor, Proto-Indo-European, is the buried root of this tree and lies outside the scope of v1.
Most handbooks date the language to roughly the second half of the first millennium BCE and place it in southern Scandinavia and on the adjoining North German Plain. Attempts to match it with a specific archaeological culture, most often the Jastorf culture, are suggestive but unproven. Language and material culture do not have to travel together.
The oldest indirect evidence is borrowing. Finnic languages took over Germanic words early enough to keep sounds and endings that the daughter languages later lost. Finnish kuningas 'king' preserves something very close to reconstructed *kuningaz, and rengas 'ring' reflects *hringaz. Roman writers such as Tacitus later recorded Germanic tribal and personal names, but no continuous text.
02
The Defining Sound Shift
What makes a language Germanic rather than just Indo-European is one systematic reshuffling of the consonants, usually called Grimm's law. Every series of Indo-European stops moved one step:
- voiceless stops became fricatives: *p *t *k → *f *θ *x
- plain voiced stops became voiceless: *b *d *g → *p *t *k
- voiced aspirates lost their aspiration: *bʰ *dʰ *gʰ → *β *ð *ɣ (later stops in many positions)
That is why Latin pater corresponds to English father, Latin decem to English ten, and Sanskrit bhrā́tar- to English brother. None of these words was borrowed. They are inherited cognates, and the consonants differ by rule, not by chance.
Grimm's law had exceptions. Gothic fadar 'father' has a d where the rule predicts þ (as in brōþar 'brother'). In 1877 Karl Verner showed that the exceptions were regular too: they depend on where the accent fell in the parent language. That result, Verner's law, was a landmark for the idea that sound change works without exceptions.
03
Stress and the Shape of Words
Indo-European had a free, mobile accent that could fall on almost any syllable. Proto-Germanic fixed it on the first syllable of the root. Verner's law must have acted before this happened, because it depends on the old accent position. That ordering is one of the firmest pieces of relative chronology in the reconstruction.
Fixed initial stress has had consequences ever since. Unstressed endings wore down generation after generation, which is part of why English and the Scandinavian languages have so little inflection left, and why alliteration (matching word-initial sounds) rather than rhyme became the backbone of early Germanic verse.
04
Daughter Branches
The traditional division has three branches: East Germanic (Gothic and its relatives), North Germanic (the Scandinavian languages) and West Germanic (English, Frisian, Dutch, German). The exact shape of the split is still debated. Many linguists think North and West Germanic went through a shared Northwest Germanic stage after the East Germanic speakers had moved away. A minority pairs Gothic with North Germanic instead. Both proposals appear under Contested Classifications below.
In the Germanic of the earliest runic inscriptions, from about the 2nd century CE, some of these splits are only just becoming visible.
Where its languages are spoken
- United States
- Germany
- United Kingdom
- Canada
- Australia
- Netherlands
- South Africa
- Sweden
26 more countries
- Austria
- Belgium
- Denmark
- Norway
- Switzerland
- Ireland
- New Zealand
- Iceland
- Finland
- Italy
- Luxembourg
- Namibia
- Faroe Islands
- Liechtenstein
- Bolivia
- France
- Mexico
- Paraguay
- Suriname
- Bulgaria
- Poland
- Portugal
- Romania
- Spain
- Tunisia
- Ukraine
Turning points
- Sound lawContested
Grimm's law
The First Germanic Consonant Shift. PIE voiceless stops became fricatives (*p → *f), voiced stops became voiceless (*d → *t), and voiced aspirates became voiced fricatives (*bʰ → *β). Latin piscis : English fish; Latin duo : English two. Its absolute date is inferred from loanwords and is disputed. Credit is disputed too: Rask described the correspondences in 1818, and Grimm's 1822 formulation made them famous.
Compare with other consonant shifts across families →
Sources (3)
- Sound law
Verner's law
Voiceless fricatives (*f *θ *x *s) became voiced (*β *ð *ɣ *z) inside a word when the preceding syllable was unaccented in the inherited accent system. It explains Gothic fadar (PIE accent after the consonant) against brōþar (accent before it), and the alternation still visible in English was : were, where *z later became r.
Compare with other weakening and voicing across families →
Sources (2)
- Sound law
Stress fixed on the root
The mobile Indo-European accent was replaced by fixed stress on the first syllable of the root. It sets up the long erosion of unstressed endings in every daughter branch.
Compare with other stress and prosody across families →
Sources (1)
- Contact
Celtic loanwords
Words tied to rank and metalwork entered from neighbouring Celtic: *rīk- 'ruler' (Gothic reiks), from Celtic rīx, and *ambahtaz 'servant' (Gothic andbahts), from Gaulish ambactos. Their Celtic sound shapes date them to a period of close contact before the Roman expansion.
Sources (2)
- Contact
Roman frontier loans
Trade along the Roman frontier brought Latin words into Germanic before the branches separated completely. Latin caupo 'petty trader' gave *kaupōn- 'to trade' (German kaufen, English cheap), vinum gave *wīna- 'wine', and Caesar gave *kaisar- (Gothic kaisar, German Kaiser).
Sources (1)
- SplitContested
Breakup into branches
As East Germanic groups moved south-east towards the Vistula and beyond, the remaining dialects diverged. Both the timing and the shape of the split are debated: a clean three-way split, a Northwest Germanic stage, or a Gotho-Nordic pairing.
Sources (2)
- WritingContested
Elder Futhark runes
A 24-letter runic alphabet appears, with early secure inscriptions such as the Vimose comb (Funen, Denmark, c. 160 CE). It was adapted from a Mediterranean alphabet, but scholars disagree whether the model was Latin or one of the North Italic alphabets. The language of the oldest inscriptions is variously labelled late Proto-Germanic, Northwest Germanic or Proto-Norse.
Sources (1)
Contested classifications
- Widely accepted
Northwest Germanic
Evidence for
North and West Germanic share changes that Gothic lacks. The long vowel *ē was lowered to ā (Gothic mēna 'moon' against Old Norse máni, Old High German māno), and the old reduplicating past tenses were remodelled (Gothic haíhait 'was called' against Old Norse hét, Old High German hiaz). The oldest runic inscriptions fit a language ancestral to both.
Evidence against
The shared features could have spread across a dialect continuum after the split instead of being inherited from one ancestor. Some, such as i-umlaut, clearly worked out differently in each branch. Scholars who favour a Gotho-Nordic pairing, or a flat three-way split, read the same evidence differently.
Sources (2)
- Largely rejected
Germanic substrate hypothesis
Evidence for
A noticeable group of common Germanic words, many to do with the sea and seafaring, has no secure Indo-European etymology. Feist also argued that a change as sweeping as Grimm's law looked like the result of a foreign population learning Indo-European.
Evidence against
Many of the supposedly isolated words have since been given plausible Indo-European etymologies. The share of words without an etymology is not unusual for an Indo-European branch. No source language has been identified, and consonant shifts of this kind can happen without language shift. Variants of the idea (for example Vennemann's proposals) survive only as minority views.
Sources (2)