English father and Latin pater (the source of French père, Spanish padre). They sound nothing alike. And yet historical linguistics does not treat them as two unrelated words that happen to share a meaning — it treats them as one and the same word, changed along two different paths. The rule behind that change is Grimm’s Law.
The clue isn’t one word — it’s a pattern
Set father next to pater and it might look like a stretch. But line up a few more pairs, and the picture changes.
| Latin | English | Sound correspondence |
|---|---|---|
| pater (father) | father | p → f |
| piscis (fish) | fish | p → f |
| pes (foot) | foot | p → f |
| tres (three) | three | t → th |
| tu (thou) | thou | t → th |
| cor (heart) | heart | c[k] → h |
| centum (hundred) | hundred | c[k] → h |
| canis (dog) | hound | c[k] → h |
A single resemblance like this could be dismissed as coincidence. Line up this many, and the story changes. Latin p always corresponds to English f; Latin t always corresponds to th; Latin c (the k sound) always corresponds to h. What Grimm demonstrated in 1822 was precisely that “always.” Not a set of isolated, ad-hoc comparisons between individual words, but a regular pattern running through the whole language.
Why does this happen at all?
As a language passes from speaker to speaker and generation to generation, its pronunciation shifts a little at a time. What matters is that this shift doesn’t happen word by word at random — the same class of sound changes the same way across the whole language. At some point in Germanic prehistory (the ancestor of English, German, Dutch, and the rest), the mouth movement for the sound p shifted, in nearly every word at once, toward the mouth movement for f. Latin and Classical Greek are the languages that didn’t go through that particular shift — which is exactly why comparing the two lets us see it as a regular divergence, rather than noise.
What “twofold” means — there’s a second layer
Grimm called this shift the “twofold” (zweifache) sound shift. The p→f, t→th, and k→h correspondences above are, in fact, only half of it. In the other half, a separate, opposite-direction correspondence happens at the same time — Latin d, for instance, becomes English t (duo “two” → two, edere “to eat” → eat). One series of sounds drops down a notch, while a different series drops down a notch of its own, in the same coordinated motion — that was Grimm’s picture of what had happened.
So why does this matter?
The value of the discovery is turning a scattering of word-resemblances into a rule you can predict from. You can now say, in advance: “if Latin has a word starting with p, the corresponding English word should start with f.” This isn’t a matter of memorizing an etymological dictionary — it’s a testable hypothesis about how language changes. You can go looking for words that fit it and check; and when you find an exception, you can ask why — expecting that the exception, too, has a reason behind it. That next step — tracking down the reason behind an exception — is the subject of a later article in this feature.
Summary
| Correspondence | Latin examples | English examples |
|---|---|---|
| p → f | pater, piscis, pes | father, fish, foot |
| t → th | tres, tu | three, thou |
| k → h | cor, centum, canis | heart, hundred, hound |
| d → t (the reverse-direction correspondence) | duo, edere | two, eat |
In a nutshell — father and pater don’t sound alike not because one of them failed to change, but because each followed its own regular path away from a sound they both once shared. The regularity is hiding precisely inside the dissimilarity.