 the journey to the Axiom of Exceptionlessness. There's a certain arbitrariness in choosing where to start any story. So I am starting with Marcus Zuerius von Boxhorn who lived from 1619 to 1653 and he's at least as reasonable a starting point as anyone. He proposed that Dutch, Greek, Latin, Persian, German, Slavic, Celtic and Baltic languages all descend from a common source and very importantly he thought that this common source was not Hebrew. At his time following biblical traditions there was a tendency to try to derive living languages from other living languages in particular from Hebrew. So in instead proposing that living languages descended from something else from a language which is no longer around he anticipated the reconstruction of proto-languages and also his choice of related languages is correct by our standards today in terms of isolating the Indo-European languages as opposed to other languages. So the next stop in our journey is Gottrib Wilhelm von Leibniz, 1646 to 1716. He's famous for many reasons including being together with Newton, one of the joint founders of Calculus. So he divided the languages of Europe into three groups depending on their work for God. You see Germanic has words for God to start with a G, Slavic words for God to start with a B and Romance words for God to start with a D. For our purposes what's important about his contribution is that he correctly identified Indo-European subgroups using some explicit criterion. Now the fact that this particular criterion got him the right answer I think was to some extent just luck but it also shows that sometimes the signal is just so strong that even using a method that would now be considered quite coarse can lead to the right answer. In other words spotting these three language families was relatively easy. So now moving on to William Jones. So William Jones is where many people begin the story of historical linguistics. He was a judge in India, a polyglot. As someone knowing Greek and Latin when he learned Sanskrit he was really struck by how similar it was to Greek and Latin and Greek and Latin and how beautiful and well-structured it is as a language. Something that I think anyone who has studied those three languages can only agree with. So he really drew attention to Sanskrit and it turns out that knowledge of Sanskrit makes the study of Indo-European much easier. So now I will just give some evidence for that. So here we're looking at Latin, Greek and Sanskrit verbs in particular the verb to bear or to carry. So I'll just read this paradigm out loud. So first person singular, pharaoh in Latin, pharaoh in Greek of course originally it would have been pronounced and then in Sanskrit. So then second person singular, phares, then third singular, and just notice the s in the second singular, the t in the third singular in both Latin and Sanskrit. The long vowel in Greek pharaoh cognate with the long vowel in Sanskrit, barami. So there are just many similarities in these paradigms. I won't talk through the other two numbers. Anyone who had learned these three languages would be naturally inclined to compare their structure and it turns out that Sanskrit provides a nice model for exploring the structure of Greek and Latin and this research agenda of using Sanskrit as a sort of key for analyzing the comparative morphology of other Indo-European languages was taken up by Franz Bopp, 1791 to 1867. So he systematically studied the comparative morphology of Indo-European languages and can really be I think considered the founder of comparative and historical linguistics. So now the next stop in our journey is Rasmus Rask, who lived from 1787 to 1832, so quite a short life, and he discovered what later came to be called Grim's Law and showed also that a Vestin was an Indo-European language, a Vestin the sacred language of the Zoroastrian religion. So now I'd like to present Grim's Law. So I'll put it together a piece at a time sort of starting from some specific components and then moving towards a more general statement. So to start with Indo-European P becomes dramatic F. I use Sanskrit and Latin, which come from quite different parts of Indo-European to exhibit this. So we have Pittar in Sanskrit, Pater in Latin, Father in English, Pada in Sanskrit, Pez in the nominative and Pettis in the genitive in Latin, and then Foot in English, Pisces in Latin, Fish in English, Purna in Sanskrit, Plenus in Latin, and Pole in English. So pretty good evidence that a P becomes an F in Germanic. Next, an Indo-European T becomes a Germanic Th, written with Theta in the IPA. So we have Tri in Sanskrit, Trares in Latin, and Tri in English, Thvam in Sanskrit, Tu in Latin, Thao in English, Tut in Sanskrit, That in English. So again, pretty good evidence that Indo-European T becomes a Germanic Th. Indo-European K becomes a Germanic Ha, pronounced Ha in English. So we have Kefale in Greek, Kaput in Latin, where the letter C makes the K sound, and Head in English, Kardia in Greek, Kors in the nominative and Kordis in the genitive in Latin and Heart in English, Kouon in Greek, Kanus in Latin and Hound in English. So that's the evidence for K turning into Ha. And now we see P changes into Fa, T changes into Tha, K changes into Ha. So altogether, we have voiceless stops becoming fricatives. Now I move a bit faster, only give one example for each specific change, and then move to these more general switches of manner of articulation. So we had stops becoming fricative, now we have voice stops becoming de-voiced. So B becomes P, for instance, Latin labium becomes English lip, D becomes T, so Sanskrit doshan, Latin dekem, we get English ken, G becomes K, Latin gelu gives us English cold. So those examples show that voice stops become de-voiced. And then we have murmured stops becoming voiced. I give Greek here, but in Greek the murmur stops were de-voiced. B becomes B, for instance, Barati in Sanskrit, Pero in Greek, and Bear in English. D becomes D, Madhu in Sanskrit, Methu in Greek, both meaning honey, and then Mead, alcoholic beverage made from honey in English. G-H, so G becomes G, Stignoti in Sanskrit, Steakene in Greek, and German Steigen, which means to climb or yeah climb step up. So putting these three sort of overall changes together, we have de becoming de, de becoming te, te becoming the. So murmur stops becoming voiced, voice stops becoming voicestops, voice stops becoming fricatives. It's as if each manner of articulation moved one over in some game of musical chairs. So that's a generalization that's extremely broad and takes in a lot of details about the development of Germanic. And now with Grims law in place, we have a tool whereby we can weed out promising but false proposals. So if you just prima facia look at English have, is it cognate with Latin Habeo? Meaning exactly the same thing. They look so similar, must be related, right? But no, English H comes from K as we've seen, whereas Latin H comes from a G-H. So they can't be cognate. And instead, Habeo is cognate with English give. And English have is cognate with Latin Capere. I think that's a nice result, right? We have a generalization. And then if we believe our generalization, we can use that generalization productively. But there are some exceptions to Grims law and that's what we'll look at next. So Carl Lautner, 1834 to 1873, who I so far have been unable to find a picture of. And if you have one, please let me know and send it to me. In an article in 1862, he catalogued all of the known exceptions to Grims law. And one point I want to make in this presentation is what a service to science that is. In the rat race of academia, there can be too much emphasis on trying to produce new results or use new methods. Well, here is a guy who just systematically collected evidence in a tidy way. And that is not an activity that gets a lot of glory, but is enormously beneficial to the progress of science. So two of his readers were then able to find patterns in these exceptions. So one was Herman Grasman, 1809 to 1877. So he solved one of the classes of exceptions to Grims law and then is also known as a great mathematician. So here is Grasman's law. First, let's look at the exceptions. So we have Grbnati in Sanskrit, cognate with grab in English. Now, the problem here is the G in Sanskrit corresponding to a G in English. We should have a G H in Sanskrit corresponding to a G in English. And then in another example, we have Bodati, cognate with Pothomai in Greek and Bid in English. And here the B in English should have a BH as its cognate in Sanskrit. So Grasman's proposal is that both in Sanskrit and Greek independently, if you had aspiration on either side of a vowel, the aspiration was lost from the consonant before the vowel. So we have Grab becoming Grab. So the G diasporating. And then Bod becomes Bod. So the B diasporating. Just to make clear, this hypothesizes that in Indo-European, everything works, right? So Indo-European had G H R BH. And then that gave us the G in Sanskrit through Grasman's law and the G in English through Grim's law. And similarly, Indo-European had something like BH U DH. And then that becomes BU DH in Sanskrit through Grasman's law. And it becomes B in English through Grim's law. There were still a few exceptions left for Carl Werner to solve. So Carl Werner was 1846 to 1896. And in an article in 1877, he solved the remaining exceptions. So Werner's law is is tricky. And it's easiest to remember using these two kinship terms. So the regular development we have is Brata in Sanskrit, Brothar in Gothic and Brother in English. Now English actually collapses the distinction that Gothic maintains. So that's why we have to turn to Gothic for the Germanic forms here. We're looking at the medial. So we have a T in Sanskrit becoming a Th in Gothic. But then in the word for father, the T in Sanskrit becomes a D in Gothic. Now what Werner noticed was that the placement of the accent was different in these two cases. So in the word for brother, the accent in Sanskrit is before the T. And then in the word for father, the accent in Sanskrit is after the T. So we can generalize that Grim's law does not apply in stops that come after unaccented syllables. Now I'll just point out that Werner's law explains what's called Gramartische Wechsel in Germanic. And I and this phenomenon is extremely marginal in English. It's really, really marginal. There are only these two examples. But I give them was and were, is and are. So you want to know what is this S in the singular, R in the plural, alternation about, we'll just notice that it's predicted by the placement of the accent in Sanskrit. So we get S changing into R when the S follows an unaccented syllable. So Werner's law is extremely impressive because it's connecting the development of constants in Germanic with the placement of accent in Sanskrit. So Werner's achievement in solving the remaining exceptions to Grim's law was extremely impressive. And two people who were impressed were Karl Brugman, 1849 to 1919, and Hermann Osthoff, 1847 to 1909. And then they founded a school of thought called the Neo-Gramarians, where they said that this result, namely, that a sound change has no exceptions is just how sound change works. So that is called the Neo-Gramarian hypothesis. Although it's probably more accurate to call it the Neo-Gramarian axiom, because it creates a research program where if we find exceptions, we have to explain them. So this axiom is that sound change is phonetically conditioned and admits of no exception. Now, apparent exceptions then can either be explained by more sound change, more refined conditions to sound change, by borrowing or by analogical developments. And I won't talk about barring or analogical developments today. Instead, we will rest content that we have completed our journey towards this Neo-Gramarian axiom.