 So having talked about neo-Gramarian sound change and how the axiom of exceptionalist came about historically I thought it would be good to talk about the other main Way that languages change internally and that is analogy So here we go Analogy and sound change with the emphasis on analogy So there are two forces internal to language sound change and analogy Sound change is a passive result of correctly learning a language It is totally regular in its application and leads to synchronic anomalies So returning to an example that I discussed in a previous presentation We have English was and were Cognate with Vasati and Vasanti in Sanskrit. So everything about the Sanskrit form is regular There is clarity of the root Vas There's movement of the accent on the root in the singular on the ending in the plural and There's clear personal endings T for the third singular and T for the third plural But nothing about the English form is regular even though it Descended from a form that was more or less the same as what we had in Sanskrit Regular sound change has obscured that regular morphology in contrast to sound change Analogy is an active result of incorrectly remembering parts of a language It is totally capricious in its application and it leads to greater regularity in terms of morphology The rest of the presentation will then Focus on analogy and I start by going through the main types. There are three main types of analogy Those are four-part analogy contamination and Analysis so taking four-part analogy first suppose I am in the middle of formulating a sentence and Wanted to say the plural of octopus Well, I might correctly remember cactus has the plural cacti, but I'm groping for the plural of octopus so In the traditional notation, I would say cactus is the cacti as Octopus is to what and then I solve for the X with octopi Because octopus is a Greek third declension noun the plural in Greek would have been octopodes So it's understandable that I that I didn't Remember that when I was formulating my sentence, of course, I could have just used the standard English plural and said octopuses But this Analogical plural Octopi is quite common So just looking at other examples of four-part analogy If I don't remember the past tense of help Maybe I will say well, what's the past tense of Google the past tense of Google is googled So the past tense of help must be helped now this example is of course facetious because help is Is an older word then Google? but the point it makes is that regular morphology Can also be explained using analogy and this is a way that we traditionalists can get around the need for Assuming that people in their head to have some kind of representation of roots and affixes We can instead just think that Analogies are made directly from one surface form to another The past tense of help was Hope and hope and in the past participle So some kind of analogy like that is how hope and hope and were replaced with help now Let's suppose we're trying to remember what the past tense of dive is And we reach for an analogy to strive we say strive is to strove as dive is to and then we solve for dove Well in this case the inherited past of dive was dived This shows us that analogy both can reinforce regular patterns and add new examples to irregular patterns Anything in the structural language is in principle available for analogical formation now taking an example from Latin Let's say I want to use the word meaning honor in the nominative and I remember the genitive is honoris and then I say well of a sister is Sororus and the nominative sister is soror so probably the nominative of honoris is Honor historically speaking. This is incorrect the nominative corresponding to the genitive honoris Was onos with with an S words not perceived as part of a pattern are generally protected from analogy so in German the way the adjective near forms its Comparative and superlative is na nea Next and that is the inherited pattern and the direct cognates of those forms in English are nigh near and Next so those three forms have been Protected from analogy in the sense that all three continue to exist in the language, but they no longer form a paradigm of plain adjective comparative and superlative of that adjective and it's precisely because they don't form that parallel that the different morphological forms have been able to continue To exist despite how divergent they become in Phenology, so the regular form is near nearer and nearest where the comparative form has Been reanalyzed as the simple form and then fresh morphology has been built on top of it so this is a nice example of where if elements of a system are Forgotten to constitute a system all of the elements or at least some of the elements you can survive independently even as the system is Reinvented analogically to take another example. We have a molten bronze where molten Exists as an adjective even though as a past participle it has been replaced with melted and Similarly, we can have the phrase cloven hoof where the past participle cloven is Preserved in this fixed expression even though The past participle Productively has been replaced with cleaved That's all I have to say about four-part analogy and now I move on to contamination when terms that present an autonomous Semantic system influence each other phonetically. That's what we call contamination So giving the most famous example, so Middle English had two adjectives borrowed from French They were male and femelle From mal and femelle in French, but in modern English the form female has been Changed to look more like the form male the form of male changing femelle into Female Because of their closely associated meaning that's what we call contamination another famous example of contamination Is found in the Indo-European numeral system So I give the words four and five in Sanskrit Greek and Latin I leave out Latin five because it has some special considerations And I also give the Mycenaean Greek version of four so that we get that nice Q in there which is the labial velar so chatur and Puncha in Sanskrit Quetoro and Pente in Greek Quattur in Latin so the we have strong evidence here for a qua Starting the word for four and a puh starting the word for five Now when we look at what happens in proto-germanic Indo-European quet war Changes into pet war in Proto-germanic before the application of Grimm's law and then that changes later into fed war for five we have pink way becoming first Pente and then Fifth the thing to focus on here is the irregular change of quae to pet in proto-germanic if We had followed the historical phonology Regularly straight from Indo-European down English should have or instead of four so what's going on here is that because four and five are semantically closely related words the P from five has been copied on to the beginning of the word for four such contaminations are extremely common in numeral systems around the world So those are two examples of contamination and just for a reminder contamination Happens in a semantic subsystem when the form of one element of that subsystem Influences the form of another element of that subsystem the third type of analogy is reanalysis so a famous example of reanalysis is that the word for for a certain kind of snake Nader in English we had said a nodder it changes to an adder and that has to do with Let's say an ambiguous surface form. So if someone says an adder You can understand it either as a natter or an adder Something Exactly parallel happened with the word napron Where a napron became an apron in a third example The word P when it was originally barred from French at an S at the end even in the singular so we had a P's and that changed into a P and that is deletion of the S because it was Reanalyzed as a plural suffix in a phrase like the P's So-called morphologically conditioned sound changes are often due to reanalysis and I will give an example of that Which is the loss of E N in past participles when the root contains a post-vocalic nasal So we say I have drunk not I have drunken So first I'll point to an author who understands this as a grammatically conditioned sound change So sealer says the designation of a specific morpheme is necessary Because some old participles have survived intact as adjectives shrunken drunken Sunken bounden These and other non-participial forms like Lenten and linen Prove that the phonological sequence alone is Insufficient to characterize the condition of loss the specific morpheme must be designated so he's saying The sound change sometimes happens and sometimes doesn't happen and it seems like then we need to Index the specific morpheme in order to know whether or not the sound change happened But long ago Otto Jesperson had given a neo-grammarian account for this same phenomenon so he says if English has nevertheless a great number of final ends it is because it has been protected by a following vowel either in the same word where the vowel has subsequently become mute or in the following word The later alternation is preserved faithfully in the two forms of the indefinite article In some words both forms survive though not used exactly in the same way Made maiden Lent Lenten drunk drunken sunk sunken So to paraphrase his argument he's saying That right after the change we had he has spoken up But he has spoke the so the occurrence of the two by forms Was entirely phonetically conditioned But it's clear that this would have been a difficult rule to learn which is why There there arose the opportunity to semantically differentiate These doublets. Thank you. That's all for now