Grimm's Law

Grimm's Law is a pattern of changes in Germanic languages. Those changes all occurred in pre-history but left evidence today.

It is Grimm's because one of the Brothers Grimm of fairytale fame discovered it (Rasmus Rask discovered it a bit earlier and Grimm built on Rask's analysis).

It is a "law" more like the law of gravity is a law than like the law against murder.

The hypothesized ancestor of the Germanic language family, called "Proto-Germanic," developed from an earlier proto-language, Proto-Indo-European. Part of the development from Proto-Indo-European to Proto-Germanic was that certain sounds "shifted."

A sound shift occurs when 1) phoneme a shifts to phoneme b, and 2) every word that includes phoneme a (in similar situations in words) stops being pronounced with phoneme a and comes to be pronounced with phoneme b.

A phoneme is the smallest sound in a language that can carry meaning: note that it does not have to carry meaning. It just has to be able to carry meaning. Consider /z/, the phoneme that makes "phonemes" different in meaning from "phoneme": the phoneme /z/ does not mean anything by itself, but it can when combined in an acceptable way with a morpheme (morpheme being the smallest unit in a language that does carry meaning).

So it's a specific phoneme that shifts over to another phoneme, and it happens universally in the language.

Note that the shifts we are going to talk about here occurred in pre-history, and only the words of Old English and other old Germanic languages show its effects. In other words, it is "hypothetical."

If you harbor any lingering doubts about the reality of Indo-European reconstruction, now is one of those times to voice those doubts. This ought to convince you that reconstruction is at least highly plausible.

In some cases, Modern English includes many examples of descendants of the same Proto-Indo-European root word in different forms, some of which underwent these shifts, others of which did not. Of those that underwent these shifts, all were old Germanic language words. Those that did not undergo those shifts came into English from non-Germanic languages (i.e. they are borrowings rather than changes). Some underwent further changes, while some did not.

There are three phases to the shifts.

1. Voiceless stops become voiceless fricatives
2. Voiced Stops become Voiceless stops
3. Voiced aspirated stops become voiced stops

  1. Voiceless stops /p/, /t/, /k/, and /kw/ become voiceless fricatives /f/, /th/, /h/ and /hw/.

PIE root
Sound Change
Modern English Derivative
Other Germanic Cognates (which show the change)
Non-Germanic Cognates (which do not show the change and reflect the  PIE) which have been borrowed into English as the words in parentheses
*ped-
foot
*p > f
foot
G fuß, Du voet, Norw/Sw vot
Gk πούς, ποδ- (>E podiatrist), L pes, ped- (> E pedal), Sanskrit pad
*peku
movable property
fee (< fehu)
ON fe
L pecu- (> E pecuniary)
*pur
fire
fire
G feuer
Gk πῦρ (> E pyr-o-mania)
*penkwe
five
five
OHG finf, ON fimmtan
Gk πέντε (> E pentagon), L quintus (> E quintet, Quentin, quintessence)
*ten
stretch
*t > th
thin

ON thunnr
(Mod. G goes to d: dunn)
Gk τείνειν (peritoneum, protasis, L tenu-is (> E tenuiroster), Skt tanu
*trei
three
three OHG thritto, G dritte Gk τρίτος (> E tritium), L tres (> E trefoil), Sanskrit trayas
*kwon
dog
*k > h
hound
G hund (E dachshund), Du hond (E keeshond) Gk κύων (> E cynic), L canis (> E canine)
*kerd
heart
heart
G herz, Sw hjärta
Greek kard-/card-, Latin cord-, Sanskrit syertse
*kannabis
hemp
hemp
OE henep
Gk κάνναβις (> E cannabis, canvas)
*kwo
relative/interrogative pronoun
*kw > hw
what
Du wat, G was, wie
L quod, quid, quam, quondam, etc. (> E quodlibet, quidnunc, quasi, quondam, etc.)
why
These changes occurred not only when the sound was at the beginning of a word: they also occurred word-internally (with significant exceptions conditioned by stress [Verner's Law]: also, two voiceless stops in a row did not both change [eight and haft would be eighth and hafth if both had changed]; and when the voiceless stop was preceded by /s/, it did not change ([e.g. spew, stand, and scold])

ONCE THAT CHANGE HAD HAPPENED, THERE WAS A 'VACUUM' LEFT: THERE WERE NO VOICELESS STOPS--NO /p/, /t/, /k/, or  /kw/ in English!

NEXT...
  1. Voiced stops /b/, /d/, /g/, and /gw/ became voiceless stops. That fills the vacuum.

PIE root
Sound Change
English Derivative Germanic Cognate (which show the change)
Non-Germanic Cognates (which reflect the PIE)
*bak
staff, cane
*b > p
peg
Du pegge
Gk βακτρον (> E bacteria)
*dekm
ten
*d > t
ten
Du tien
L decem (> E decemvir, decimate)

seat

L sedes (>E sedentary)





sweet

L suade- (>E dissuade, persuade, suasive)

tame

L domus (> E domestic)
*gel
cold
*g > k
cold
Du koud, G kalt
L gelu (> E gelid)
*gwei
live
*gw > kw

quick
Du kwiek
Gk βίος, L vivus (> E vivid)
*gwou
cow
cow
Du koe, Old Frisian
Skt gaus, gaum, go (>E cow), L bos (> E Bovine)
*gwen
woman
queen

Gk γυνη,
Again, these changes also occurred word-internally.

ONCE THAT CHANGE HAD HAPPENED, THERE WAS A 'VACUUM' LEFT: THERE WERE NO VOICED STOPS--no /b/, /d/, /g/, or /gw/!

NEXT...
  1. Voiced aspirated stops became plain voiced stops. That fills the vacuum.

PIE root
Sound Change
English Derivative Germanic Cognate (which show the change)
Non-Germanic Cognates (which reflect the PIE)
*bha
*bh > b
beacon, beckon
mDu boenen "scour, scrub"
Gk φῶς (> E phosphorus)
*bhrater
brother
G bruder, Du broeder
Gk φρατήρ (> English phratry), L frater (> E fraternal)
*bhreu
brew
OHG briuwan, ON bruggin
Gk φρεαρ (> E phreatic), L fervere (> E fervent)
*dheu
*dh > d
dead
ON dauthr, Gothic dauths, G tot

*dhwer
door
OHG tor, ON dyrr, Gothic daur
Gk θύρα, L fores
*ghans
goose
*gh > g
goose
gander
Frisian goes, G gans, Du gans

L anser (>E anserine, merg-anser), Gk χήν (>E chen-o-pod)
*gwhibh
*gwh > gw
wife
Frisian wif, G weib

Again, the changes occurred word-internally as well.

And nowadays there are no voiced aspirated stops left in English.