Today’s post is brought to you by the letter G

It’s time for the HLC with our very special guest, Proto-Germanic! Yaaay!

Ah, English spelling. That prickly, convoluted briar patch that, like an obscure Lewis Carroll poem, often falls just a little too shy of making sense. Or does it?

It wasn’t always like this. English spelling actually used to be pretty phonetic. People would just write down what they heard or said.1 Then, the printing press was introduced. Books and pamphlets began to be mass produced, literacy levels rose, and spelling began to be standardized. At the same time, English continued to move through some fairly dramatic shifts in pronunciation. The language moved on as the spellings froze.

Throughout the years, people have occasionally called for reforms in English spelling. Like that time in the early 20th century when Andrew Carnegie, Melvil Dewey, Mark Twain, Theodore Roosevelt, et. al. colluded to “improve” some of the more confusing orthographic practices of English. Personally, this linguist is glad such efforts have by and large failed.

Sure, you could look at English spellings and tear at your hair at the monumental insanity of it all. But I like to think of our spellings more as fossils preserving the dinosaur footprints of earlier pronunciations. Granted, sometimes the footprints are from five different species, all overlapping, and there’s, like, a leaf thrown in.

Where are they all going?!

Let’s take, for example, the letter <g>2 and its many possible pronunciations.

First on the menu is the classic [g], a sturdy stop found in words like grow, good gravy, and GIF. This dish originates in the Proto-Germanic (PGmc) voiced velar fricative /ɣ/3. (Refresh your memory on our phonological mumbo-jumbo here.) This velar fricative had a bit of an identity crisis during Old English (OE)4, spurred on by hanging out with sounds all over the mouth.

“But what we found out is that each one of us is a front vowel…and a back vowel…and a palatal approximant…an affricate…and a voiced velar stop…Does that answer your question?”

Around front vowels (such bad influences—triggering umlaut wasn’t enough for them?), it became [j], as in year, from OE ġēar. Between back vowels (the big bullies), it became [w], as in to draw, from OE dragan5. At the end of words, it lost its voicing and became [x] (the sound in loch), as in our own dear Edinburgh (whose pronunciation has since changed again). Ah, but before back vowels, and when backed up by sonorants like [ɹ], it held its ground a little better and became our trusty [g].

As you may have noticed, a lot of the sounds that came from /ɣ/ are no longer spelled with <g>. Alas. We’ll come back to how Edinburgh wound up with an <h> in a minute.

But first, there was another sound that came from PGmc /ɣ/. Old English had something going on called gemination. Sometimes, it would take a consonant and double its pronunciation. Like the <kk> in bookkeeper. Bookkeeper is just fun to say, but these long consonants were actually important back in OE. The wheretos and whyfors of gemination are another story, but just like how /ɣ/ became [j], the geminate /ɣɣ/ was pulled forward and dressed in new clothes as the affricate [d͡ʒ], like in bridge and edge, from OE bryċg and eċg.

Gemination didn’t get around much. It was pretty much restricted to the middle of words. When mushy, unstressed endings began to fall off, the leftovers of gemination found themselves at the end of words, but a little nudge was needed before [d͡ʒ] found its way to the prime word-initial position. Later on in Middle English, the language ran around borrowing far more than a cup of sugar from its neighbor across the Channel. As English stuffed its pockets with French vocabulary, it found a few French sounds slipped down in among the lint. One of those was Old French’s own [d͡ʒ], which on the Continent was simplifying to [ʒ]6 (the <s> sound in measure). This [ʒ] sound didn’t exist in English yet. Our forefathers looked at it, said “nope,” and went on pronouncing it [d͡ʒ]. Thus we get words like juice, paving the way for later words like giraffe and GIF.

This is a GIF. Or is it a GIF? I mock you with my scholarly neutrality.

It was only later, after the end of Middle English, that /ʒ/ was added to the English phoneme inventory, retaining its identity in loanwords like garage and prestige. It’s worth noting, however, that these words also have accepted pronunciations with [d͡ʒ].

Alright, so what about the <gh> in Edinburgh? It turns out there’s another sound responsible for the unpaid overtime of the letter <g>. Meet the sound /h/. In Middle English, Anglo-Norman scribes from France introduced a lot of new spellings, including <gh> for /h/. The <h> part of the <gh> digraph was probably a diacritic meant to indicate a fricative sound. Remember that by this time, the old <g> didn’t really represent a fricative anymore. In words like Edinburgh, the [x] from /ɣ/ had merged with the [x] version of /h/, so it is from /h/ that we get our <gh> spellings. Over time, these [h] and [x] pronunciations weakened and disappeared completely, bequeathing us their spelling to baffle future spelling bee contestants. We have them to thank for bright starry nights, the wind blowing in the high boughs of the trees. But before these sounds went, they left us one last piece to complete our <g> puzzle: after back vowels, sometimes [x] was reanalyzed as [f]. We’ve all been there, right? Your parents say something one way, but you completely mishear them and spend the rest of your life pronouncing it a different way. I mean, did you know the line in the Christmas song is actually colly7 birds, not calling birds? Now imagine that on a language-wide scale. I’m glad for the [f]s. They make laughing more fun, although sometimes convincing your phone not to mis-autocorrect these words can be rough. Had enough? Okay, I’ll stop.

The point of all this isn’t really about the spellings. Just look at all these beautiful sound changes! And this barely scratches the surface. A lot of the big sound changes that warrant fancy names seem to be all about vowels, but as <g> can attest, consonants have fun, too.8 Speaking of big, fancy vowel changes, get your tickets now because next week, Sabina’s going to talk about one of the most famous and most dramatically named: the Great English Vowel Shift.

Notes

1 It wasn’t a perfect system, though. Sometimes, a single scribe would spell the same word several different ways in the same document. Was this reflecting variations in utterances? An inability to decide which letter represented which sound? Transmission errors through copying down someone else’s writing? Who knows.
2 As far as the letter itself goes, the Anglo-Saxons actually used a slightly different symbol known as the insular g. The letter we use today was borrowed from the French during Middle English and is known as the Carolingian g.
3 It’s the voiced version of the sound at the end of Scottish loch. It can be heard today in the Dutch pronunciation of wagon.
4 Refresh yourself on the periods of English here.
5 Actually, draw, drag, and draught/draft are cognates. Knowledge, am I right?
6 This is actually one of my favorite phones. I’m a linguist. I’m allowed to have favorite phones.
7 Because they’re black like coal. And my heart.
8 Admittedly debatable and unnecessarily anthropomorphizing, but we’re already in this thing pretty deep.

Once upon a time…

Once upon a time, there were two brothers who very much enjoyed stories. They travelled their country looking for folk tales, each one darker and grimmer than the last… There was no happily ever after in sight and, though their stories have changed much since, the original tales are still found out there for those brave enough to seek them…

Prepare yourselves, my dears, because this… this is the story of the brothers Grimm.

*

Or not! Actually, it is the story of one of the brothers: Jacob Grimm. And it won’t be grim in the least but full of fun linguistic facts!

Today, we’ll be talking about what is known as the First Germanic Sound Shift, Rask’s Rule or, most commonly, Grimm’s Law.

Riccardo touched upon this topic in last week’s post on the comparative method, a method that was pretty much born with this particular observation. The first to notice the correspondence that would eventually become Grimm’s Law was Friedrich Schlegel, a German philologist, in 1806. Rasmus Rask, a Danish philologist, extended the ‘rule’ to to other PIE languages in 1818 and, eventually, Grimm included German in his book Deutsche Grammatik, published in 1822.

Now, they noticed a regular sound change that affected certain Proto-Indo-European (PIE) consonants. They also noticed that this particular sound change only affected the Germanic languages, e.g. German, Dutch, English, Swedish, etc.

But what is it?

Well, Grimm’s Law describes how certain PIE consonants developed in Proto-Germanic, particularly early Germanic stops and fricatives. Now, you might want to refresh your memory on phonological terminology before continuing, but there can be said to be three parts of the chain shift that is Grimm’s law:

  1. PIE voiceless stops became voiceless fricatives
  2. PIE voiced stops became voiceless stops
  3. PIE voiced aspirated stops became voiced stops or fricatives.

That might be a bit abstract but it basically works like this:

PIE PGmc¹ PIE PGmc PIE PGmc
p > f b > p bh > b
t > θ d > t dh > d
k > x g > k gh > g
> > ghʷ >

 

Consider these words in Latin, English and Swedish and compare them to their PIE root:

 

PIE² Latin English Swedish
*ped- pēs foot fot
*dwo- duo two två
*genu- genū knee³ knä

Now, why would English and Swedish have <f>, <t> and <k> where PIE and Latin have <p>, <d> and <g>?

Well, because English and Swedish, being Germanic languages, underwent Grimm’s Law and thus changed the PIE sound */p/, */d/ and */g/ to /f/, /t/ and /k/ respectively. Latin, on the other hand, is an Italic language and didn’t undergo this change, thus keeping the sounds of PIE (or at least approximately, though exactly how close these sounds are is a bit difficult to say with certainty).  

Why would this happen, you might wonder? What would make one sound shift to become another sound?

Well, we don’t really know exactly how it started or why. It might be what is called a ‘pull chain’, meaning that one sound shifts, leading to a ‘gap’ in the phonological values of the language. As a result, another sound shifts to fill that gap and a third sound shifts to fit the gap of the second one and so on and so forth.

But, it is also possible that it worked the other way around, meaning that one sound started to shift and basically pushed another sound out of its place, thereby leading to a chain shift. This is called a push chain.

But as to how such a chain started? Well, that part is still kind of shrouded in mystery. Perhaps two sounds became too similar to each other and became difficult to distinguish from each other, forcing a shift? We might never know.

What we do know, however, is that Grimm’s Law did affect all Germanic languages, leading to a distinction between that language family and its PIE-derived sisters.

But there are also a good number of exceptions from this rule. For example:

Why does PIE *bʰréh₂tēr (“brother”) become Proto-Germanic *brōþēr but PIE *ph₂tḗr (“father”) became Proto-Germanic *fadēr?

In ‘brother’, the development follows Grimm’s Law, i.e.  t > þ, but in ‘father’ it does not. Instead of the, by Grimm’s law, expected development, i.e. t > þ, the Proto-Germanic word developed t > d. Why is that?

Well, cue Karl Verner; a Danish linguist who in 1875 formulated what is now known as Verner’s Law, an addition, if you will, to Grimm’s Law. Verner’s Law explains such occurrences as ‘father’, showing that voiceless fricatives, e.g. *f, *s, *þ, when immediately following an unstressed syllable in the same word, underwent voicing and becomes fricatives, e.g. *β, *z,*ð

Now, you might be thinking that this is all very interesting but why is it important? ‘cause I can pretty much promise you, that if there is anything the budding historical linguist is aware of, it is Grimm’s Law.

Well, while it is fascinating in its own right, its discovery showed us something much greater than we had ever thought possible before: that sound change is a regular phenomenon, not a random process affecting only some words.

This discovery not only set historical phonology apart as its own field of study but also means that we can predict and understand phonological developments, a discovery that cleared the field for the comparative method.

And without the comparative method, of course, our field of inquiry would be so much poorer as we would largely be unable to properly understand the relationship between languages and the historical developments of those languages.

And wouldn’t we all be a lot poorer for that lack of understanding?

So, next time you watch Cinderella, Little Red Riding Hood or Hansel and Gretel, remember that Jacob Grimm not only provided you with these stories but helped design the most used, and important, method in historical linguistics to this day. Not a bad contribution, right?

Join us next week when our awesome magician Riccardo is back! This time, he’ll be talking about the magic of umlaut and ablaut, so if you’ve ever wondered why it’s ‘mouse’ but ‘mice’ but not ‘house’ and ‘hice’ you definitely don’t want to miss it.

 

Notes and sources

¹ PGmc is a common abbreviation for Proto-Germanic

² All the PIE roots can be found by a simple google search. These are taken from the Online Etymology Dictionary found here: https://www.etymonline.com/. Have fun!

³ Remember now that while the <k> in modern English ‘knee’ is silent today, it was pronounced in earlier stages of English.

*The little pic is from http://tentcampinghq.com/camping-articles/how-to-tell-scary-campfire-stories-2/

**For those who wants to know more about Grimm’s Law, most (if not all) introductory textbooks on linguistics deals with the subject at least a little bit. This particular illustration is from Millward, C.M. A Biography of the English Language.  Ft. Worth: Harcourt, 1996. Pg. 63 but a similar one can be found in pretty much any textbook. Particularly recommended is Lyle Campbell’s Historical Linguistics (3rd ed., 2012) which deals with most things historical linguisticky with great attention to detail and plenty of examples (so it’s recommended generally, not only for this particular sound change). 

Is English a Romance language? On language families and relationships

Today, I’m going to talk about language families! When I say this, I believe that most of you will have, on some level, an intuitive hunch about what I mean. If we were to compare a couple of common words found in, for example, Spanish and Italian, we would find that they are often very similar or, in some cases, even identical. Take a look:

Spanish Italian English translation
vivir vivere live
boca bocca mouth
tu you


Similarly, if we were to look at Swedish, Danish and Norwegian:

Swedish Danish Norwegian English translation
leva leve leve live
mun mund munn mouth
du du du you


You see the similarities? Now, why is that, you might wonder. Well, because they are related!

In the linguistic world, related languages are languages that have so much in common that we cannot claim that it is merely due to extensive contact and/or borrowing. These languages, we say, are so similar that there can be no other reasonable explanation than that they descend from a common source: a mother language, as it were. In the case of Spanish and Italian, the mother is Latin, while in the case of Swedish, Danish and Norwegian, the language is Old Norse.

Now, it would be convenient if it stopped there, wouldn’t it? But, of course, it doesn’t. Like any family, the mother also has a mother and other relatives, like siblings and cousins. Old Norse, for example, has its own sisters: Old High German, Old Frisian, Old English, etc., which all share the same mother: Proto-Germanic. This is the Germanic language family.

Spanish and Italian also have sisters: French, Portuguese, Romanian, etc., and their common mother is Latin. This is the Romance language family, deriving from Vulgar Latin. But, of course, Latin has its own sisters, for example Umbrian and Oscan, and together with its sisters, Latin forms the Italic language family.

Does it feel a bit confusing? Well, that’s understandable and I’m going to kick it up a notch by adding that the Italic language family, with languages like Spanish and Italian, and the Germanic language family, with languages like Swedish and Danish, actually have the same mother: Proto-Indo-European (or just Indo-European).

The mother in this case is veeeery old, and we actually don’t have any kind of evidence of how it looked! Indo-European is a reconstructed language, more commonly known as a proto-language (as you may have noticed, we call the mother of the Germanic family Proto-Germanic, meaning that it is also a reconstructed language). It has never been heard, never been recorded and no one speaks it. Then how the heck do we know anything about it, right? Well, that has to do with something called the comparative method, which we’ll explain in another post.  

Like human families, language families can be represented in the form of a family tree:*

Clear? Well, hate to tell you this, but this is an extremely simplified version using only examples from these two subfamilies. The “real” Indo-European language family tree looks somewhat more like this:1

You’re kinda hating me right now, aren’t you?


As you can see by the tree above, some languages that you might never expect are actually related. Let’s take as an example Standardised Hindi and German. Here are some common words in both languages:

German Hindi English translation
Mädchen लड़की (ladakee) girl
Hallo नमस्ते (namaste) hello
Hunger भूख (bhookh) hunger


Looking at these words, it is unlikely that you would draw the conclusion that the two languages are related. Looking at the language tree, however, you can see that linguists have concluded they are. Now, you’re probably staring at your screen going “whaaaat?” but, indeed, they are both descendants of Indo-European and are therefore related.

While Indo-European is clearly a large group of languages, it is not the only one (or even the largest). Looking a bit closer at the Indo-European language family, you will notice that languages such as Mandarin and Finnish are not included. These belong to other families, in this case the Sino-Tibetan and Finno-Ugric (or Uralic, depending on your definition) language families respectively.

All in all, there are approximately 130 language families in the world today. Some are related, some are not, just like we are. The largest family is the Niger-Congo language family, having (as recorded in 2009) 1,532 languages belonging to it. (Indo-European comes in a poor 4th place with approximately 439 languages.)2

So, looking at languages is kinda like looking at your own family tree: every mother will have a mother (or father, if you want, but traditionally, linguists call them mothers and daughters). Some branches will have siblings, cousins, second cousins and so on. Some will look nothing like their relatives (or, well, little anyway) and some will be strikingly similar. That’s just the way families work, right?

So, now, we’ve reached a point where I can answer the question in the title: Is English a Romance language?

While this is a much-debated question (do a google search and see for yourself), the simple answer is: no, it’s not. At least, not to a linguist. Now, you might be sitting at home, getting more and more confused because a lot of English vocabulary can be traced back to Latin (the word ‘vocabulary’ being one of those words, actually).

But when linguists say that a language is a Romance language, we are referring to the relationship illustrated in the tree structure, i.e. the language has Latin as its mother. English, then, despite having borrowed a substantial part of its vocabulary from Latin (and later from the Latin language French), it is not in itself a daughter of Latin. English is a daughter of Proto-Germanic, thus, it is a Germanic language.

However, Latin and Proto-Germanic are both daughters of Indo-European. Latin and English are therefore clearly related, but the relationship is more like that of a beloved aunt rather than a mother (if, you know, the beloved aunt refused to recognise you as a person unless you imitated her).

At the end of the day, languages are like any other family: some relationships are strong, some are weak, some are close, some are not.

Tune in next week when Riccardo will delve into another branch of language families: constructed languages.

Notes and sources

*The structure employed here, showing languages as families in family trees, has long been criticized for simply not showing a lot of information like contact-situations, dialect continuums and when the languages were spoken. It has, however, been used to show the beginning student that some languages are related to each other and how they are related in a way that is easy and comprehensible. The Historical Linguist Channel does, however, recognise this criticism and would be happy to discuss it in a separate post or through personal communication.  

1Provided by Ancient History Encyclopedia (Published on 19th of January, 2013).  https://www.ancient.eu/image/1028/

2Lewis, M. Paul (ed.), 2009. Ethnologue: Languages of the World, Sixteenth edition. Dallas, Texas: SIL International. Online version: http://www.ethnologue.com/16. (Family index is reached through http://www.ethnologue.com/16/family_index/).