A Quick Explainer of Grimm's Law
A (roughly) 1000-word rundown of phonology's most famous discovery
Background
If you’ve read this blog, you’ve heard the phrase “Grimm’s Law.” A few months ago, I contrasted Grimm’s Law with a similar shift consonant shift in Armenian. It was too long, and it was somehow both too academic and too cute. Thus, I wanted to write a new piece that explained Grimm’s Law while removing all the other crap. Here is that piece.
Let’s start with some basics. It’s not a law. It’s not the linguistic version of F=MA. The Romance languages, for example, have not undergone anything that looks like Grimm’s Law. Instead, think of Grimm’s Law as a historical event, like the Great Vowel Shift in Early Modern English. To clear up this confusion, some refer to Grimm’s Law as the “First Germanic Consonant Shift.” That would have been a better name, sure, but no one calls it that. If you read anything about historical linguistics, it will probably say “Grimm’s Law.” It’s a bad name, so just substitute it with “Grimm’s Historical Event” in your head. One last thing before you Google it: yes, this is the same Grimm who cataloged fairy tales.
Two more terms for you: Proto-Indo-European is the parent language of the Germanic, Italic, Greek, Slavic, Indo-Aryan, and various other languages. It was spoken in the Russian steppes around 6,000 years ago. Proto-Germanic is the parent language of English, Dutch, German, Icelandic, the Nordic languages, and various dead East-Germanic languages like Gothic. It was spoken around 2,500 years ago in northern Germany and southern Denmark. Note the “Proto” in front of each name. Unlike Sanskrit or Latin, no one wrote down these ancient languages. Instead, they’re pieced together through observational evidence and linguistic theory. Think of Neil DeGrasse Tyson explaining how the Earth was formed. Neither he nor any other physicist was alive back then. They’re modeling the Earth’s creation by observing its present state and comparing it to planets that are mid-formation in other parts of the universe.
A lot of time passed between Proto-Indo-European and Proto-Germanic. No one went to sleep speaking one and woke up speaking the other. Think of Proto-Germanic as a dialect of Proto-Indo-European that slowly drifted into a separate language, the same way Latin slowly drifted into Italian. We can’t piece together every change that occurred during this drifting. Linguists theorize the start (Proto-Indo-European) and the end (Proto-Germanic), and they infer the changes from those points.
A couple more caveats. Let’s start with the notion of a “consonant shift.” This refers to speakers replacing all (or, at least, a large number) of one consonant sound with another. You’re probably familiar with the fact that we say something like “hafta” instead of “have to.” The v→f change in “hafta” is not a consonant shift. That’s one isolated incident of assimilation, where sounds morph to fit alongside their neighbors. A consonant shift would involve all the “v” sounds in the whole language (or all the v sounds before certain vowel sounds, for example) shifting into “f” sounds.
These shifts occur gradually. In both England and North America, we’re in the middle of a consonant shift involving the “th” sounds. You can buy “Norf London” in the UK, and you’re probably familiar with phrases like “dat ass” in the US and Canada. I don’t have any numbers on this, but imagine that roughly 5% of Americans speak with the “dat ass” accent. Maybe that moves to 7% next generation and 9% for the generation after that. By the time we hit the 24th century, that hits 100% and people start wondering why everyone has a lisp in those really old movies.
Last item: when explaining Grimm’s Law, I always use examples from modern English. No, Proto-Germanic didn’t sound much like 21st-century English. However, the consonants are close enough for the comparisons to work. You might also be aware of different consonant sounds in Dutch and German. That’s because these languages had their own, later shifts that have nothing to do with Grimm’s Law. The Dutch word "goed” (cognate of English “good”) begins with /γ/, which kind of sounds like someone clearing her throat. That’s due to a consonant change that’s unique to Dutch. We’ve kept the original “g” in English. German speakers refer to the number “10” as Zehn (pronounced “tsayn”). Again, they changed, not us. The word began with a “t” in Proto-Germanic. The differences between Dutch, English, German, or any other Germanic language have nothing to do with Grimm’s Law.
Enough caveats, let’s get to Grimm’s Law.
Grimm’s Law
We’re done with caveats, but we still need some definitions. Stop consonants block the airflow before releasing it in a sudden burst. English’s stop consonants are b, p, t, d, k, and g. Fricative consonants create turbulence in the throat. In English, these are v, f, the two “th” sounds, z, s, “s” sound in “measure,” sh, and h. Each of these come in pairs. Put your hand on your throat while saying “vvvvvvvvvvvvvvv” and “ffffffffffffffffffffffff.” Your voice box vibrates during the former, making it a voiced consonant. The other sound is unvoiced. Finally, put your hand in front of your mouth and say “Pop.” You’ll notice that you release air during the first p, but not the second. We refer to that first “p” as breathy or aspirated, and we represent that breathiness with a super-script h: pʰ.
Grimm’s Law concerns the stop consonants. In the standard model of Proto-Indo-European, these occurred in three series. They looked like this:
There’s also a second, less accepted reconstruction of Proto-Indo-European known as the Glottalic Theory:
This one removes the aspirated/breathy series and adds a group of ejective consonants. I recommend listening to some ejective consonants here, because, as the old saying goes, an MP3 is worth a thousand words.
The glottalic theory provides a more parsimonious explanation for some Indo-European sound changes. However, no ancient or modern Indo-European language contains an ejective consonant, giving the theory an “assume a can opener” vibe. For our purposes, we’ll focus on the standard model. If true, those are the stop consonants you would have heard in Eurasian steppes about 6,000 years ago. Yet, when your time machine stops in 500 BC Denmark, you would have heard this:
That is Grimm’s Law. All the aspirated and voiced stop consonants lost their aspiration. You won’t hear a bʰ, dʰ, or gʰ today. All the voiced stop consonants switched to their unvoiced partner. All the unvoiced stops became fricative consonants.
It’s easy to see this change when we contrast our native words with their Latin equivalents. Among series one consonants, we can contrast trio and three, pesc and fish, or cornucopia and horn. For series two, think of duo and two or labia and lip.