This is the nerdiest hobby I'm really incredibly interested in. Like, most of my _other_ nerd hobbies have broken into the mainstream somehow. But conlangs are so detail-focused and have such a strange utility that every time I bring it up in public people wince. It's amazing.
When I bring it up, people think that I’m making some kind of code of English, and when I bring up writing systems, people ask stuff like, “How do you write [insert letter of english alphabet]?”
The conlang I put on hold after seeing "How NOT to make a language" did some fun things with irregularity. It had verbs that marked singular vs. plural, and in the protolanguage they formed the plural by reduplicating the first syllable. But after sound changes that deleted vowels all over the place and subsequent consonant cluster reduction, the two forms could look wildly different. It had paradigms like singular "ran" vs. plural "nengan", singular "khezhan" vs. plural "khetaran", etc.
6:10 I have several problems with this frame. First of all, it implies that the English word “have” comes from the Latin “habeo”, when the two words are actually completely unrelated, and second of all, it follows no clear pattern, so I’m not even sure of the point you’re trying to make with it. On top of implying that “habeo” is Spanish for “I have” when it was inherited as “he”, you used the example of its use as a future tense marker, when for the French and English you used the perfect auxiliary. French also fused forms of “avoir” into its infinitives to express future tense. English just doesn’t belong in the graph.
0:38 делать means "to do", not "to say". Same with the past tense feminine conjugation. Also, as an aside nitpick, идти and шла here are imperfect verbs, so it's more like "to be going by foot" rather than "to go on foot".
@@dj3us The thing is that it doesn't directly translate in English. Идти really only applies to short distances, since there's no universal word like "to go" for movement in Russian. That's why it's usually written as "to go on foot".
@@dj3us Technically we can go long distances on foot, no shit, but generally the use of идти applies to short distances when talking of "to go" and ехать applies to farther distances because generally we walk short distances and drive long ones. Am I wrong?
идти is from PIE *h₁éyti 'to go' and шла is from ходить :) Slavic languages tend to lose D, e.g. in Polish we don't say serdce but serce, because it's hard to say D between R and C, but we can say easly they cognates: serdeczny, serducho, środek. The same do East Slavic languages with e.g. endind -DŁO, when they use -ŁO, e.g. *rydlo --> rylo, *gъrdlo --> gorlo. In 'new' words you have the tendency to writing it but not speaking so e.g. идти́ /ɪˈtʲːi/ or се́рдце /ˈsʲert͡sə/ So we have feminine form шла and masculine form шёл, which in Polish we still speak with D, so szła (f) and szedł (m). ш (sz) is palatalized х (ch), e.g. mucha (муха) NOM --> musze (мухе) LOC/PREP (In Polish locative has a strong palatalizing ending, e.g. pora --> porze; las --> lesie [ś]; byt --> bycie [ć], noga --> nodze, ręka --> ręce etc.), or *duxъ --> *duša. If you change sz [š] in szedł [šedł] to ch [x] you have xedł --> *xodilъ from *xoditь. The development of the word-forming category of the aspect meant that the Perfectum and the Imperfectum (past tenses) became unnecessary, so Proto-Slavic lost them. The so-called PS’ Imperfectum had been a new creation and had been used to describe ongoing (durative) and repeated (iterative) actions in the past. So идти is durative and ходить is iterative verb, the same like e.g. пить and пива́ть. There is a question: Can х- naturally appear before иди in a new word, to became ходи? For some reason it have more sense to me than that *xoditь is from *sědětь :)
6:08 It's obvious he made a mistake with the text but I don't know what the text is supposed to be. On one hand there's French in infinite and the "ai" doesn't match it, and on the other there's "I have" conjugated.
@@fernandobanda5734 "ai" *is* the present tense, 1st person singular conjugation of "avoir" (and it's also the auxiliary for French perfect) But you're right, it's unclear what the goal of that text was.
Minor mistakes I found, following me reälizing their presence via one of the other comments: 1) The Indo-Europeän chart you show is not correct. Armenian and Greek are not in the same family, and Irish is not a romance language. 2) Latin _habeō_ does not precede English _have._ The English form is from Proto-Germanic _habjaną_ and is etymologically unrelated. 3) The Proto-Indo-Europeän root for "to be" is _"h₁es-",_ not _"es-."_ The laryngeäl accounts for several Sanskrit verb forms and the fact Greek has an initial vowel in all forms of the word, among other things. 4) The Proto-Indo-Europeän root for "good" is _"bʰed-",_ not _"bʰad-"._ Although there are arguments that at some stage or another that one or both of PIE's vowels was some value of /a~ɒ/, PIE doesn't have any phonemes written with , unless you disregard the laryngeäls (a very bad ideä). Germanic _bataz_ comes from an o-grade form _(bʰodos>botos>batas>bataz)_ and reflects the merger of post-PIE /a/ (from e+h₂) and /o/ (from either original o or e+h₃). Good video, though!
@@novvain495 If you read attentively, he seems to use them for differencing the digram /i:/ or /ɛː/ (non exhaustive) from hiatuses or pseudo-hiatuses /iæ/ or /iə/. Quite ingenious I must say.
Irish is almost certainly included with the Italic languages because many Indo-Europeanists in the modern day accept an "Italo-Celtic" grouping between the two obviously-closely-related subgroups.
The chart accepts the Armeno-Hellenic and Italo-Hellenic sub-groupings, i.e. that those branches share a relatively more recent unattested common ancestor
Interesting how the plural of "person" is "people", but both those words can be further pluralized (via the regular suffixing of -s) into "persons" and "peoples" in some contexts to evoke different meanings, despite the non-affixed forms being considered the same "word" by native speakers.
Suppletion happens quite often in Ancient Greek. For example, the Present of the verb "to see" is ὁρῶ (horō), and you'd expect its aorist (simple past) to be ὥρησα (hōrēsa) because word-initially the vowel becomes long and verbs that have an accent on the ō ending gain an -ēsa suffix. Instead you have εἶδον (eidon) which is completely unrelated and forms like an Imperfect (with the on ending). Turns out that εἶδον is in fact the Imperfect of an obsolete verb that in an archaic form of Greek meant "to be informed), εἶδον meant I was informed but slowly became "I saw", because how else would you be informed back then? Another verb took εἶδον's place as "to be informed" and we're left with a type that makes no sense!
Two words...THANK YOU! I am creating my first conlang and I was really wondering about how could some things be so irregular and how could I implement this in my conlang, and you came to save the day! Yours is truly the best conlang channel!
I’m one language, I have the word for Eue, “Ēl”. But as it was a word used quite often as the people were shepherds, it’s collective form (in fact the most common form of “Ēl”) became “Rhym”, instead of the regular “Ēlym” you would expect. It’s because the older form was “Iel”, with “Ielhym” as the collective form, but eventually the h was lost in -hym, but in “Ielhym”, it mutated into “Iel̥ym”, and [l̥] became [r̥], becoming “Ierhym”. (Also the diphthong “ie” became “ē”.) And the “Ē” at the beginning of “Ērhym” got lost because of how often it was said, becoming “Rhym”. Also three forms of irregularity are: 1) Past tense verbs can often have a different vowel indicating it’s past tensey ness because a proto form of marking the past tense was to shorten a long vowel. And countless sound changes have really messed with it. 2) Imperfect tense verbs used to be indicated by reduplicating the first syllable. But sound changes have made that practically irrecognizable. 3) A proto form of marking the plural was by umlaut.
And don't forget to apply some overcorrection from time to time, like adding an ending to a word that already is in that form but uses an older production instead of using the base form and adding the ending to that. Or applying some rule that doesn't belong to a word just because it looks like it fits the class the rule applies to, e.g. applying the old-regular form to a regular word that rhymes with an irregular one---and then applying the regular suffix in addition to that. (e.g. greet -> grate (like eat-ate) -> grated)
Well, yes, Japanese does have two entirely irregular verbs, and those are "kuru (to come)" and "suru (to do, make)." The copula "desu (sort of 'to be')" isn't really a verb, so much as it is either its own part of speech, or it's really just a conglomerate of old particles and other historical oddities. Also, other verbs, like "iku (to go)," have a small number of irregular parts, including those verbs which have suppletive forms for honorific and humble speech.
Ablaut. Not Umlaut. Ablaut. In no group of germanic verb does umlaut carry even a hint of "pastness". Consider weorðan, whose present tense first person form is "weorðe", and the past "wearð", while umlaut would have produced *wierð. The actual vowel alteration goes in the exact opposite direction. And just btw, English to have and latin habeo (I dunno how to type a macron) are not cognates. habban/hafian go all the way back in (Old) English They couldn't have been borrowed and are usually reconstructed as unrelated in PIE as well. Also >to be >Implying we have need of such base, filthy words, used by commoners
Your videos are really awesome! I have one request for a little feature focus though - Stress/timing/mora systems. I sort of understand the rough theory around them but am having a little difficulty just linking that to how it actually sounds in some cases.
To add to what you said of the "to be" form, don't forget that some languages don't use this. I speak Russian and we don't use "to be" form even though it does actually exist in Slavic languages and the words for it exist in Russian as well despite not being used.
"делать" is a word for "to do". "to say" in russian is "сказать" /skə'zatʲ/, "(she) said" is "сказала" /skə'zaɫə/. and tbh it's pretty regular in russian.
Oh and yes, and some languages keep both words even though suppletion happens. Swedish has Gammal, Äldre Äldst, for old older and oldest, but it also has ålder åldra åldrats, and gamla gammlare gammlast. And I have a hunch that gamalaz and aldaz have some common link in PIE,
A nitpick around 6:06: with the possible exception of رأى, all of the Arabic verbs listed are actually better analyzed as falling into an alternate regular conjugation, not a true irregular pattern. For example, كان (to be) conjugates exactly like داب (to melt or dissolve). You even have a paired example listed: أخذ and أكل follow the same conjugation pattern, as would all verbs of the standard verb whose root starts with the glottal stop.
Fun bit missing from 7:36 for Spanish: the fact that "fui" apparently kept the meaning of "was" as well, so the same forms mean "was" and "went" depending on context.
one video a few months ago was talking about the most irregular verbs in English. it talked more a linguist that was able to find a math equation to show use vs regularness. this also ment we can predict what verb is likely to become regular next. It likely going to be Wed, so don't be surprised to hear "i we are newly weded" or "should these two not be weded".
so fascinating; believe it or not; because of this video; now I understand that the most common words being irregular is not a secret plot to mess with language learners heads; that is a theory that is just plausible enough that knowing how it actually arises is nessecary to debunk that
Really great overview of how it happens in natural languages. I struggle sometimes with how to move forward with things I know I want to do/achieve with my conlangs because I just don’t understand how natural languages do it, and therefore my toolbox is empty to replicate it naturalistically. I appreciate you filling my toolbox for irregularity, sir!
Hey, I've just recently discovered your channel, watched a few videos and... I really like WHAT you feature in this videos, there's some very interesting stuff, but I don't like HOW you do it - I get lost after just a few minutes, trying to track both what you're saying and what's on the screen, as those two things sometimes are only related, but not interacting. I'd really like it if you explained it all using the visual part as an illustration to what you're saying, not additional in-depth information - and for that, a bit longer, yet more in-depth (in voiceover!) videos would be nice. Also, a bit more time to see what example did you just show, as now I have to pause every now and then to read it in time. I just thought I'd share it with you, as I'm usually catching new concepts quick, so there might be others having that problem with getting your content.
I realize that all known languages have irregularities not all these ireregualarities are found in verbs, Hungarian is said to have no irregular verbs but rather irregular nouns
Turkish only has one, disputably even, because I've heard and seen the regular form being used instead of irregular one, often. So, I'm sceptical to the claim that every language have some, or at least one. Seems like my language can lose the only one it possesses in the future.
@@soton4010 The genitive case of the noun warer, "su" is "suyun", meaning of (the) water, belonging to water, whereas according to our current rules it should've been "sunun". There's only one suffix for the genitive case for the all nouns, and it has been so since a very long time. And that's the sole irregularity in the whole language. As I've said, I've heard and seen "sunun" too.
Zaccari Jarman, what kind irregularity do you mean? There are some irregular plural nouns in Hungarian. A fun fact about the verb 'to be': It has a future conjugation (no other verb does) and the future conjugation has its own, separate past form.
It's interesting to think on forms of irregularity as I go into reworking my main conlang. One idea I really want to stick with in my new version is the use of the location cases for states of being, e.g. "child-nominative adult-allative be-continuous" = "the child is growing up" or "friend-nominative pron-1st-sg-genitive disease-ablative be-continuous" = "my friend is recovering from illness." I'm now thinking that a suppletion of forms of the verb "to become" slipped into the continuous forms of the verb "to be" after these usages became commonplace, as the case meanings made the specific verb a bit redundant, but people were still used to using "become" in that context.
As a student of the Polish language, I'd be interested to know where it got its perfective/imperfective verbs from, and specifically why there are so many different rules for deriving one from the other. For example some verbs add z as a prefix to go from imperfective to perfectice, as in _robić/zrobić, jeść/zjeść,_ some verbs add 'po' as a perfix, as in _myśleć/pomyśleć, słuchać/posłuchać,_ some verbs add 'yw' as an infix to go the other way, from perfective to imperfective, as in _zachować/zachowywać,_ some verbs change some of the vowels, as in _przeszkadzać/przeszkodzić,_ some change some of the letters in a seemingly arbitrary way, as in _dotykać/dotknąć, otwierać/otworzyć, umierać/umrzeć,_ and some change completely, such as _brać/wziąć, widzieć/zobaczyć, mówić/powiedzieć._ But all these pairs are understood as imperfective/perfective counterparts.
Pining for the fjords All the prefixes like z-, po-, za- have subtle nuances which would take a long time to explain, and the same could be said for the verb suffixes like -a-, -ywa-, -ną-, -i-, -ie-... However the way all these elements turned into a perfective/imperfective system is quite natural and logical. otwierać/otworzyć is just the e/o ablaut which has been around since PIE, while the other two are originally vowel length. tuk/tūk -> tk/tyk, mir/mīr -> mr/mir -> mr/mier (i->je before “r” is a specifically Polish shift). As for why some suffixes trigger a lengthening of the root vowel, I can’t tell you. Suppletion was already covered in the video, zobaczyć was originally the counterpart of baczyć, and powiedzieć of powiadać, but other imperfective verbs with the same meanings (widzieć, mówić) largely replaced them, but baczyć and powiadać can still be used in more limited contexts with specific nuances.
ywa is not an infix. ...ać and ...ić - that is not just change in vowel, that's a removing of -a- suffix, and I don't know what is -i-. The pairs where the words change completely are the cases of suppletion
7:35 Note that Standard Italian officially has no form from "ire", but a lot of regional dialects have. But I don't know how much this is due to the Spanish dominion of past centuries and how much has instead remained from Latin directly.
@@wk_vylion, I thought so. Thanks for the confirmation about affixes and European loanwords. Language change is natural, and no committee can prevent change even if it succeeds in slowing it. I knew that doctoral study in linguistics would come in handy some day.
Another thing I would have added is foreign languages. I feel as if you are planning on developing a single conlang then this is unnecessary but if you are worldbuilding and a couple of cultures are close enough to trade, invade or merge with eachother then radical changes will occur in your language too, maybe making some words irregular.
It's always interparadigmatic analogy so if there's regular ablaut in a set of words or even just one common word people will use this for "regular" words to make them "irregular". Some linguists think this might be how the Semitic languages got their triconsonantal roots
I was thinking back to your last episode realized that you be interesting to see a video combining the two concepts and seeing what a alien language you might sound like using speculative biology to see the kind of sounds other creatures might make and incorporate into a language
Thanks for the videos, such a great and very interesting one. Just remember that habeo is 'to have' in Latin, not in Spanish! That'd be 'haber'! Nontheless awesome vid
Raumanoetro has the adjective [go], meaning "equal" in the masculine singular. It becomes [jekwi] in the masculine plural, [kɛwa] in the feminine, and [kɛwɛ] in the masculine plural. Edit: this was just the result of sound change.
pfap [p̪ap] (child) proto-Bolen pfapki [p̪apˈki] (child-ᴘʟᴜ) proto-Bolen pfap [p̪ap] (child) SB pfapki [p̪akʲ] (child-ᴘʟᴜ) SB In the orthography the word for child is regular, but in speech it isn't. (I used what I call an "orthographical romanisation", where instead of writing the sounds you write a representation of an alphabet that doesn't have unicode, you write in a representation of it in the latin alphabet)
What about other grammatical irregularities that come from relics of the old language? English has one particular phrase construction that illustrates. "Here comes the bus." The "here" of that sentence isn't going anywhere. It is "the bus" that is the subject of the sentence even though it is in the object slot of that sentence. Similar phrases of "There goes ..." are also useful. Historically, "good" german or old english would place the verb second in the sentence, but as English lost its noun declensions for nominative and accusative cases, we gained fixed word order instead. But we retain this little phrase that make sense for us, but grammatically is odd. Do you have any recommendations for how to incorporate this kind of irregularity into a language and if there are other languages that have these little artifacts like English in them?
Not sure if this is a good idea or not, but what about just, NOT conjugating verbs, like, at all? If I recall correctly, Mandarin does this, but I'm not 100% sure.
I mean, sort of. Morphologically, yeah, Mandarin has no verb conjugations, but can one really say that the different particles and modals around the verbs are not conjugations, AT ALL? I mean, there might not be any person or number agreement, but I would think of the various particles as some kind of conjugation.
@@Sovairu Sorry for the late reply. But yeah, I was speaking morphologically, yeah. I'm sorry if I came off the wrong way. I'm not sure if I'll include particles and such in mine or not, but I do know that I won't be modifying verbs in mine.
@@Yoshimaster96smwc Oh, no, your comment didn't come off the WRONG way; it just seemed a bit vague or incomplete. I mean, if you want the form of your verbs to never change, and have other separate words in the sentence do all the heavy lifting for person, number, TAM, voice, evidentiality, conjunction, or anything else, that's fine; go wild! I just think of it as sort of a gray area with saying that such a system does or does not have conjugations at all. However, to be fair to you, after I went back to the almighty Wiki to check on conjugations, it seems that they are pretty much understood to be different inflections on the verb, instead of separate words ( en.wikipedia.org/wiki/Grammatical_conjugation ). So, yeah, if you want your word for the verb eat to only ever look like "et" or something, then go ahead. It's perfectly fine!
all irregularity can ultimately be analysed as a very baroque "regular" system with lots of ifthenotherwises - somebody did this for Latin nouns. I forget his name, but I do recall that the nominative case gave him the most trouble.
In a few hundred years, I predict that the infinitive form of "can" (which English doesn't have right now) will be replaced by some ghastly contraction of "to-be-able-to", which will confuse language learners for centuries to come.
I think we should make defective verbs complete again. To can, to will, to shall, to may, to mote (the obsolete present form of _must)_ and all the others I forgot.
If I wanted to reform an existing language instead of creating a new one, how would I do so? I’m looking for a way to make the English language more consistent, without actually changing any words or grammar, just by making a script that better fits the language. Any tips on where I should start? I’m thinking of taking a system like Hangul and transforming it into an Alpha-Syllabary thing, with the many irregularities of English in line.
Though irregularity is most often the wrong term. The "irregular" verbs of modern English still follow a regular rule. Only and are truly irregular as their paradigms have different roots which can't be predicted.
The past participle of go comes from another root, indeed. More specifically, from wend, which means turn, return, and follows the same conjugation pattern as send.
Really, a regular rule predicts all of drink->drank, write->wrote, freeze->froze, wreak->wrought, make->made, split->split, sleep->slept! (and doesn't also predict think->thank, bite->bote, wheeze->whoze, sneak->snought, take->tade, sit->sit, and beep->bept!) I'd love to hear about this magical rule!
@@AzrgExplorers many of the irregulars you are listing can be predicted through ablaut relations, although English pronunciation and spelling can randomize it a bit. True irregulars would be verbs shifting to different roots, or taking suffixes other than -ed, unless there is a way to determine them through phonology.
@@alternateperson6600 If you have to study the etymology to even halfway predict the forms of verbs, what's the point in calling them "not truly irregular"? A language learner still has to just memorize them!
@@AzrgExplorers nobody has even said a thing about etymology, although some verbs which were once "irregular" (shifted to preterit through Germanic Umlaut) might have lost their ablaut inflection in the transition to Modern English. Sadly enough, it seems like no one has gone so far as to compile some sort of material with the ablaut relations for verbs which take in umlaut to change tense. It would greatly help English learners, hell, even native speakers, who think "irregular" verbs are completely a hit or miss dilemma. But because English is, somewhat, conservative on etymology, some ablaut relations haven't changed at all, even if the radical vowel has shown changes on pronunciation, so that Old English can serve as a valuable reference to compile such material from scratch.
Look at Navajo, for one, as it is exclusively prefixing (at least on verbs), and a lot of weird things happen when those prefixes start piling up. Or rather, do the same things as with any other affixes, sounds, or languages. This sounds flippant, but, really it's true. No matter what kind of affixes or morpheme combinations you have in a word, they will usually count as the one, whole word, and diachronic and synchronic sound changes will result based on the whole word.
About the same way. Instead of final vowel loss, you could have some syncope in pre-stress syllables. However, if it was me doing it, I'd introduce special sound changes just for intervocalic consonants, most likely some combination of progressive palatalization and vanilla lenition. If clusters are allowed, I'd probably also have some unique outcomes for certain clusters.
It's hard learning all of this in my native language, but it's disheartening to realise that it's even harder in English even when it's simplified to oblivion XD
One irregularity that really threw me off when I first saw it was the Breton word for "do" because its infinitive "ober" is a loanword from Latin "opera" but the conjugated forms ("graet" among others) are native verb forms similar to those in Welsh and Cornish. I know many languages, such as English with "be", combine different word roots for basic verbs but this the first time I've seen a language go as far as to borrow one of the roots used in their formation.
I have a question. I'm currently working on my first serious conlang, and I was wondering if it's naturalistic for a language to differentiate between alienable and inalienable adjectives and have them be in different positions in the sentence. For example "Taly teitrsy Fa" means, "I'm 16" and "16" comes after "Taly" (To be) because age can change easily. However, "Taly Fa batrda" means "I'm a man" and "A man" comes after the first person pronoun; "Fa" because gender is an innate thing. My logic is that the "alienable" descriptions are explaining how the subject is IN THAT MOMENT and how easily changeable that state is. And the "inalienable" descriptions is how the subject HAS BEEN and how it is nearly impossible to change that description. Thanks!
I'm particularly fond of the fossilized conjugations process. Like how in English alot of multisylabic words that end in -us become -i and and -a becomes -ae and both are pronounced "ai"
I think that the English simple past tense used to be the imperfect but the its role as the imperfect was replaced by a past form of *to be* and the present participle. I believe that *to be* still has irregular past tense forms is due to it once being an imperfect.
That's an awesome video but I just want to point out that, out of all the Arabic verbs you listed, only ليس (not to be) is truly irregular in that it can only conjugate for past. All the other verbs in your list have "regular irregularity".
The verb "To be" in Swedish is: To be: Att vara I am: Jag är You are: Du är He/She/It is: Han/Hon/Den/Det är We are: Vi är You are: Ni är They are: De är So basically, it's always "är", although I have no idea how "vara" changed to "är".
I’m sorry if it is too much of a hassle, but could you provide the sources you read to understand the methods of grammar creation in your “how to” video series? I’m in the process of making a conlang now but I’m having some trouble with Adpositions in particular. Plus I’d like to be able to read the information myself for more reference. Thanks for your help!
Well, this is a bit weird. Ancient Greek too has this thing with good Ἀγαθὸς (Agathos, good) Βελτίων (Beltiōn, better) Βέλτιστος (Beltistos, best) The suffixes are nothing extraordinary but the adjective is considered irregular. Also, for some reason, the verb to be in Ancient Greek is very regular, although from a rare category of verbs ending in -μι (mi), which is the oldest IE suffix, while normal verbs end in -ω (ō) even in modern Greek. This is surprisingly the most inexplicable suffix of Greek, where as the other person conjugations came directly from PIE. Apparently, it comes from Ἐγώ (Egō, meaning I) and it's made to comply with its suffix.
I hope to see a video on the topic of mixed conlangs aka fictional creoles/pinjins, there is not much information on this topic in the conlanging community.
Oh, the copula. So irregular that Genki literally teaches it as if it weren't a verb, but some other class of word altogether (though still at the same time).
My conlang has irregular verbs that only slightly deviate from the regular pattern (as in they are one or two missing or transposed letters away from being perfectly regular) and only in some of their conjugations (as in just two or three out of the six tenses are irregular). I'm not sure how realistic this is, and I'm wondering if I should try to kick the weirdness of these verbs up a notch.
This doesn't sound natural, thoguh again you didn't say much. Irregularities that stand out are more likely to remain than those that are a nudge away from being regular. Also, irregularities tend to be regular themselves, e.g. most irregular verbs will follow one of only a few subpatterns.
@@Релёкс84 There is a set of what I've dubbed "alternative verbs" that follow a different pattern and are regular amongst themselves. These are much more distinct from the common regulars, but I wasn't considering them truly irregular because there's so many of them and the pattern they follow is entirely predictable if you know which ones are alts.
is there anything about a natural language that can 'force' high regularity? the best i can think of is being a pidgin/creole, which i heard tend to be especially grammatically simple. i could also imagine that speakers could enjoy the elegance or regularity, but maybe that'S wishful thinking. needless to say i'm more interested in engineered languages than naturalistic ones. :P
You can 'force' high regularity by using a language reform. Language reform can occur in two ways: simplification and purification. Simplification can reduce irregularities in a language while purification is another way to introduce even more irregularities.
The Academy, and planned languages (Ausbau, mostly). That's basically when politics prevents the natural evolution. That happens quite often, and even more for the written form (the best example I know of is castillan spanish that became almost phonemic with a single, intense reform).
If your language developed all its inflections recently, they're likely to be highly regular, maybe with a few predictable phonological rules like the "tundu" example.
@@senesterium yeah, language plannung seems to usually have a fairly auhoritarian nature. i'd like something more bottom-uppy, but maybe you'd need a society of linuists/conlangers for that..
Different sound changes are the first step. (they start of as dialects) Example sound changes: In Estonian vowels were dropped from the ends of words: keeli > keel In Finnish long mid vowels became diphthongs: keeli > kieli
I present to you the forms of "to-be" in arthumonabuz: thoý [θ̠oy] (imperfect) oý [oy] (perfect) dni [dni] (future) the perfect "oý" used to be the verb for "to-rest" in the protolang but over time became the copulae, the imperfect: "thoý" uses the old past continuous, verbs in arthumonabuz usually don't have a future but to be is the exeption, "dni" is just an old verb for "to-grow"
This is the nerdiest hobby I'm really incredibly interested in. Like, most of my _other_ nerd hobbies have broken into the mainstream somehow. But conlangs are so detail-focused and have such a strange utility that every time I bring it up in public people wince. It's amazing.
It's like, "i make my own languages"
"Why?"
"as a hobby"
"What? Isnt that really hard?"
"yeah, but it's fun"
"...What?"
"but how will you get people to speak it??!"
"How do you say [insert phrase with a weird grammatical construction that you don't know how to replicate in your conlang]?"
Tbh I wince talking about it...
When I bring it up, people think that I’m making some kind of code of English, and when I bring up writing systems, people ask stuff like, “How do you write [insert letter of english alphabet]?”
I was going to make my own language. Then I realized that Latin has all of the qualities I'm looking for in a conlang. So I'm learning Latin.
I won't lie. That's hilarious.
I won't lie. I'm doing the same.
What would those be?
Neque ego mentiar: ridiculum non video.
Have you tried Esperanto perchance? ;)
7:17 In Dutch, 'wenden' means 'to turn', wouldn't be surprised if it has the same roots as 'went'
Yes it does. Also, 'went' was the past tense of 'to wend', which means 'to turn', but is now obsolete, and 'went' is now the past tense of 'to go'.
As well as in German :)
Scandinavian languages too
Germanic, people
In English, we still have 'to wind'
The conlang I put on hold after seeing "How NOT to make a language" did some fun things with irregularity. It had verbs that marked singular vs. plural, and in the protolanguage they formed the plural by reduplicating the first syllable. But after sound changes that deleted vowels all over the place and subsequent consonant cluster reduction, the two forms could look wildly different. It had paradigms like singular "ran" vs. plural "nengan", singular "khezhan" vs. plural "khetaran", etc.
your language sounds dope! honestly it sounded like my own, since i use -an as simple past
6:10 I have several problems with this frame. First of all, it implies that the English word “have” comes from the Latin “habeo”, when the two words are actually completely unrelated, and second of all, it follows no clear pattern, so I’m not even sure of the point you’re trying to make with it. On top of implying that “habeo” is Spanish for “I have” when it was inherited as “he”, you used the example of its use as a future tense marker, when for the French and English you used the perfect auxiliary. French also fused forms of “avoir” into its infinitives to express future tense. English just doesn’t belong in the graph.
0:38 делать means "to do", not "to say". Same with the past tense feminine conjugation. Also, as an aside nitpick, идти and шла here are imperfect verbs, so it's more like "to be going by foot" rather than "to go on foot".
“To go on foot” is “идти пешком”…
“Идти” is just “to go”…
@@dj3us The thing is that it doesn't directly translate in English. Идти really only applies to short distances, since there's no universal word like "to go" for movement in Russian. That's why it's usually written as "to go on foot".
@@mikzin630
А чё это я не могу длинную дистанцию пройти?
@@dj3us Technically we can go long distances on foot, no shit, but generally the use of идти applies to short distances when talking of "to go" and ехать applies to farther distances because generally we walk short distances and drive long ones. Am I wrong?
идти is from PIE *h₁éyti 'to go'
and
шла is from ходить :) Slavic languages tend to lose D, e.g. in Polish we don't say serdce but serce, because it's hard to say D between R and C, but we can say easly they cognates: serdeczny, serducho, środek. The same do East Slavic languages with e.g. endind -DŁO, when they use -ŁO, e.g. *rydlo --> rylo, *gъrdlo --> gorlo. In 'new' words you have the tendency to writing it but not speaking so e.g. идти́ /ɪˈtʲːi/ or се́рдце /ˈsʲert͡sə/
So we have feminine form шла and masculine form шёл, which in Polish we still speak with D, so szła (f) and szedł (m). ш (sz) is palatalized х (ch), e.g. mucha (муха) NOM --> musze (мухе) LOC/PREP (In Polish locative has a strong palatalizing ending, e.g. pora --> porze; las --> lesie [ś]; byt --> bycie [ć], noga --> nodze, ręka --> ręce etc.), or *duxъ --> *duša. If you change sz [š] in szedł [šedł] to ch [x] you have xedł --> *xodilъ from *xoditь.
The development of the word-forming category of the aspect meant that the Perfectum and the Imperfectum (past tenses) became unnecessary, so Proto-Slavic lost them. The so-called PS’ Imperfectum had been a new creation and had been used to describe ongoing (durative) and repeated (iterative) actions in the past.
So идти is durative and ходить is iterative verb, the same like e.g. пить and пива́ть. There is a question: Can х- naturally appear before иди in a new word, to became ходи? For some reason it have more sense to me than that *xoditь is from *sědětь :)
Ironic that "good" in PIE was "bad".
_bhad,_ but yeah
no, wait that's still not right; it's _bhed-_
Where did the "a" come from? PIE doesn't have an "a" phoneme...
@@Copyright_Infringement it could be *bheh2d- with a laryngeal
@@gaveferia1421 That would take some explaining, given भद्र exists
@@Copyright_Infringement oh it is *bhed- but Germanic used the o-stem, giving PGmc *bataz
@@gaveferia1421 Yup! There's even a long grade reflex (bōtō)
6:00
Spanish "Haber" conjugates to "he" in the first person. Not habeo. If only.
I think that's the wrong timecode
6:08
It's obvious he made a mistake with the text but I don't know what the text is supposed to be. On one hand there's French in infinite and the "ai" doesn't match it, and on the other there's "I have" conjugated.
@@fernandobanda5734 "ai" *is* the present tense, 1st person singular conjugation of "avoir" (and it's also the auxiliary for French perfect)
But you're right, it's unclear what the goal of that text was.
Crap, that was supposed to be "he" for Spanish. Must have somehow copied the wrong text.
Habeo is in Latin, if I remember well
Minor mistakes I found, following me reälizing their presence via one of the other comments:
1) The Indo-Europeän chart you show is not correct. Armenian and Greek are not in the same family, and Irish is not a romance language.
2) Latin _habeō_ does not precede English _have._ The English form is from Proto-Germanic _habjaną_ and is etymologically unrelated.
3) The Proto-Indo-Europeän root for "to be" is _"h₁es-",_ not _"es-."_ The laryngeäl accounts for several Sanskrit verb forms and the fact Greek has an initial vowel in all forms of the word, among other things.
4) The Proto-Indo-Europeän root for "good" is _"bʰed-",_ not _"bʰad-"._ Although there are arguments that at some stage or another that one or both of PIE's vowels was some value of /a~ɒ/, PIE doesn't have any phonemes written with , unless you disregard the laryngeäls (a very bad ideä). Germanic _bataz_ comes from an o-grade form _(bʰodos>botos>batas>bataz)_ and reflects the merger of post-PIE /a/ (from e+h₂) and /o/ (from either original o or e+h₃).
Good video, though!
I like the random Äs popping along your comment.
@@novvain495 If you read attentively, he seems to use them for differencing the digram /i:/ or /ɛː/ (non exhaustive) from hiatuses or pseudo-hiatuses /iæ/ or /iə/. Quite ingenious I must say.
I thought the Irish thing was odd too. Wikipedia says it’s Celtic.
Irish is almost certainly included with the Italic languages because many Indo-Europeanists in the modern day accept an "Italo-Celtic" grouping between the two obviously-closely-related subgroups.
The chart accepts the Armeno-Hellenic and Italo-Hellenic sub-groupings, i.e. that those branches share a relatively more recent unattested common ancestor
Interesting how the plural of "person" is "people", but both those words can be further pluralized (via the regular suffixing of -s) into "persons" and "peoples" in some contexts to evoke different meanings, despite the non-affixed forms being considered the same "word" by native speakers.
Has your video quality gone up? It seems so.
Suppletion happens quite often in Ancient Greek. For example, the Present of the verb "to see" is ὁρῶ (horō), and you'd expect its aorist (simple past) to be ὥρησα (hōrēsa) because word-initially the vowel becomes long and verbs that have an accent on the ō ending gain an -ēsa suffix. Instead you have εἶδον (eidon) which is completely unrelated and forms like an Imperfect (with the on ending). Turns out that εἶδον is in fact the Imperfect of an obsolete verb that in an archaic form of Greek meant "to be informed), εἶδον meant I was informed but slowly became "I saw", because how else would you be informed back then? Another verb took εἶδον's place as "to be informed" and we're left with a type that makes no sense!
Amazing video. Suppletion in particular is a fascinating concept.
Two words...THANK YOU! I am creating my first conlang and I was really wondering about how could some things be so irregular and how could I implement this in my conlang, and you came to save the day! Yours is truly the best conlang channel!
I’m one language, I have the word for Eue, “Ēl”. But as it was a word used quite often as the people were shepherds, it’s collective form (in fact the most common form of “Ēl”) became “Rhym”, instead of the regular “Ēlym” you would expect.
It’s because the older form was “Iel”, with “Ielhym” as the collective form, but eventually the h was lost in -hym, but in “Ielhym”, it mutated into “Iel̥ym”, and [l̥] became [r̥], becoming “Ierhym”. (Also the diphthong “ie” became “ē”.) And the “Ē” at the beginning of “Ērhym” got lost because of how often it was said, becoming “Rhym”.
Also three forms of irregularity are:
1) Past tense verbs can often have a different vowel indicating it’s past tensey ness because a proto form of marking the past tense was to shorten a long vowel. And countless sound changes have really messed with it.
2) Imperfect tense verbs used to be indicated by reduplicating the first syllable. But sound changes have made that practically irrecognizable.
3) A proto form of marking the plural was by umlaut.
And don't forget to apply some overcorrection from time to time, like adding an ending to a word that already is in that form but uses an older production instead of using the base form and adding the ending to that. Or applying some rule that doesn't belong to a word just because it looks like it fits the class the rule applies to, e.g. applying the old-regular form to a regular word that rhymes with an irregular one---and then applying the regular suffix in addition to that. (e.g. greet -> grate (like eat-ate) -> grated)
I like how Japanese has only 2 irregular verbs. "suru" is often called irregular, but it's conjugation never changes.
Well, yes, Japanese does have two entirely irregular verbs, and those are "kuru (to come)" and "suru (to do, make)." The copula "desu (sort of 'to be')" isn't really a verb, so much as it is either its own part of speech, or it's really just a conglomerate of old particles and other historical oddities. Also, other verbs, like "iku (to go)," have a small number of irregular parts, including those verbs which have suppletive forms for honorific and humble speech.
"its conjugation never changes" is irrelevant. As the video clearly state, the irregularity is relative to the main pattern, not to itself…
@@senesterium Yeah. What does "its conjugation never changes" even mean?
6:10 "Habeo" is not a word in Spanish. The verb "haber" does exist but the conjugation for first person singular is "he"
Now I you're just making things up. Every time I watch your videos I wonder how I can speak my native language while knowing nothing about it.
Seme
Ablaut. Not Umlaut. Ablaut.
In no group of germanic verb does umlaut carry even a hint of "pastness".
Consider weorðan, whose present tense first person form is "weorðe", and the past "wearð", while umlaut would have produced *wierð. The actual vowel alteration goes in the exact opposite direction.
And just btw, English to have and latin habeo (I dunno how to type a macron) are not cognates. habban/hafian go all the way back in (Old) English They couldn't have been borrowed and are usually reconstructed as unrelated in PIE as well.
Also
>to be
>Implying we have need of such base, filthy words, used by commoners
>
I wish I could like your videos twice! So good!
Happy pi day! Thank you for all your conlanging episodes. You are amazing
7:25 lol, "bad" used to mean "good"
that's why it's better and best and not gooder and goodest
this may be the best conlangery video i've seen. well done.
Your videos are really awesome! I have one request for a little feature focus though - Stress/timing/mora systems. I sort of understand the rough theory around them but am having a little difficulty just linking that to how it actually sounds in some cases.
To add to what you said of the "to be" form, don't forget that some languages don't use this. I speak Russian and we don't use "to be" form even though it does actually exist in Slavic languages and the words for it exist in Russian as well despite not being used.
Yes, that is I wanted to say! Russian to be used so rare that it is completely regular.
0:42 делать is ‘to do’
To say is ‘сказать’
Another amazing video from Biblaridion! I'm just about to start work on my language family so this is really handy to keep in mind!
@6:20 According to Fowler's Modern English Usage (1927), the word "art" has made the word "beest" archaic in the second person singular.
"делать" is a word for "to do". "to say" in russian is "сказать" /skə'zatʲ/, "(she) said" is "сказала" /skə'zaɫə/. and tbh it's pretty regular in russian.
Well it *was* shown as an example of a regular verb
@@Mercure250example with wrong translations is bad
Делать is to do not to say and about other. These videos are amazing
Oh and yes, and some languages keep both words even though suppletion happens.
Swedish has Gammal, Äldre Äldst, for old older and oldest, but it also has ålder åldra åldrats, and gamla gammlare gammlast.
And I have a hunch that gamalaz and aldaz have some common link in PIE,
interesting, in Danish we have Gammel(old), Ældre(older), and Ældst(oldest). Very similar to Swedish.
A nitpick around 6:06: with the possible exception of رأى, all of the Arabic verbs listed are actually better analyzed as falling into an alternate regular conjugation, not a true irregular pattern. For example, كان (to be) conjugates exactly like داب (to melt or dissolve). You even have a paired example listed: أخذ and أكل follow the same conjugation pattern, as would all verbs of the standard verb whose root starts with the glottal stop.
Fun bit missing from 7:36 for Spanish: the fact that "fui" apparently kept the meaning of "was" as well, so the same forms mean "was" and "went" depending on context.
one video a few months ago was talking about the most irregular verbs in English. it talked more a linguist that was able to find a math equation to show use vs regularness. this also ment we can predict what verb is likely to become regular next. It likely going to be Wed, so don't be surprised to hear "i we are newly weded" or "should these two not be weded".
so fascinating; believe it or not; because of this video; now I understand that the most common words being irregular is not a secret plot to mess with language learners heads; that is a theory that is just plausible enough that knowing how it actually arises is nessecary to debunk that
Really great overview of how it happens in natural languages. I struggle sometimes with how to move forward with things I know I want to do/achieve with my conlangs because I just don’t understand how natural languages do it, and therefore my toolbox is empty to replicate it naturalistically. I appreciate you filling my toolbox for irregularity, sir!
Just as a suggestion, any chance of a video on initial consonant mutation (such as occurs in the Celtic languages)?
Of course you would find a way to insert some Navajo talk in there...
Never change, Biblaridion, never change!
Hey, I've just recently discovered your channel, watched a few videos and... I really like WHAT you feature in this videos, there's some very interesting stuff, but I don't like HOW you do it - I get lost after just a few minutes, trying to track both what you're saying and what's on the screen, as those two things sometimes are only related, but not interacting. I'd really like it if you explained it all using the visual part as an illustration to what you're saying, not additional in-depth information - and for that, a bit longer, yet more in-depth (in voiceover!) videos would be nice. Also, a bit more time to see what example did you just show, as now I have to pause every now and then to read it in time.
I just thought I'd share it with you, as I'm usually catching new concepts quick, so there might be others having that problem with getting your content.
I realize that all known languages have irregularities not all these ireregualarities are found in verbs, Hungarian is said to have no irregular verbs but rather irregular nouns
Turkish only has one, disputably even, because I've heard and seen the regular form being used instead of irregular one, often.
So, I'm sceptical to the claim that every language have some, or at least one. Seems like my language can lose the only one it possesses in the future.
@@Balequalm what's irregular in Turkish that not verbs?
@@soton4010 The genitive case of the noun warer, "su" is "suyun", meaning of (the) water, belonging to water, whereas according to our current rules it should've been "sunun". There's only one suffix for the genitive case for the all nouns, and it has been so since a very long time.
And that's the sole irregularity in the whole language.
As I've said, I've heard and seen "sunun" too.
Zaccari Jarman, what kind irregularity do you mean? There are some irregular plural nouns in Hungarian.
A fun fact about the verb 'to be': It has a future conjugation (no other verb does) and the future conjugation has its own, separate past form.
Which makes some sense with such an extensive case marking system.
My obsessions,
Computer science
Genetics
Neurology
Computer engineering
Mechanical engineering
Electrical engineering
maths
Psychology
Chemistry
And finally linguistics/conlangs
It's interesting to think on forms of irregularity as I go into reworking my main conlang. One idea I really want to stick with in my new version is the use of the location cases for states of being, e.g. "child-nominative adult-allative be-continuous" = "the child is growing up" or "friend-nominative pron-1st-sg-genitive disease-ablative be-continuous" = "my friend is recovering from illness." I'm now thinking that a suppletion of forms of the verb "to become" slipped into the continuous forms of the verb "to be" after these usages became commonplace, as the case meanings made the specific verb a bit redundant, but people were still used to using "become" in that context.
You're stuff has really helped with my conlanging and I can't wait to see more.
As a student of the Polish language, I'd be interested to know where it got its perfective/imperfective verbs from, and specifically why there are so many different rules for deriving one from the other. For example some verbs add z as a prefix to go from imperfective to perfectice, as in _robić/zrobić, jeść/zjeść,_ some verbs add 'po' as a perfix, as in _myśleć/pomyśleć, słuchać/posłuchać,_ some verbs add 'yw' as an infix to go the other way, from perfective to imperfective, as in _zachować/zachowywać,_ some verbs change some of the vowels, as in _przeszkadzać/przeszkodzić,_ some change some of the letters in a seemingly arbitrary way, as in _dotykać/dotknąć, otwierać/otworzyć, umierać/umrzeć,_ and some change completely, such as _brać/wziąć, widzieć/zobaczyć, mówić/powiedzieć._ But all these pairs are understood as imperfective/perfective counterparts.
Pining for the fjords All the prefixes like z-, po-, za- have subtle nuances which would take a long time to explain, and the same could be said for the verb suffixes like -a-, -ywa-, -ną-, -i-, -ie-... However the way all these elements turned into a perfective/imperfective system is quite natural and logical.
otwierać/otworzyć is just the e/o ablaut which has been around since PIE, while the other two are originally vowel length. tuk/tūk -> tk/tyk, mir/mīr -> mr/mir -> mr/mier (i->je before “r” is a specifically Polish shift). As for why some suffixes trigger a lengthening of the root vowel, I can’t tell you.
Suppletion was already covered in the video, zobaczyć was originally the counterpart of baczyć, and powiedzieć of powiadać, but other imperfective verbs with the same meanings (widzieć, mówić) largely replaced them, but baczyć and powiadać can still be used in more limited contexts with specific nuances.
ywa is not an infix.
...ać and ...ić - that is not just change in vowel, that's a removing of -a- suffix, and I don't know what is -i-. The pairs where the words change completely are the cases of suppletion
7:35 Note that Standard Italian officially has no form from "ire", but a lot of regional dialects have. But I don't know how much this is due to the Spanish dominion of past centuries and how much has instead remained from Latin directly.
Does this mean to suggest that Esperanto may develop irregular verbs in the future if it develops naturally?
@@wk_vylion, I thought so. Thanks for the confirmation about affixes and European loanwords. Language change is natural, and no committee can prevent change even if it succeeds in slowing it. I knew that doctoral study in linguistics would come in handy some day.
Did you mean to say "ablaut" at 4:34?
Say->said
See->saw
Yee->haw
Yeet -> Yaught
@@gabor6259 actually it's yote
Jelq-jelqed
@@Nehauonget that regular verb out of here
Another thing I would have added is foreign languages.
I feel as if you are planning on developing a single conlang then this is unnecessary but if you are worldbuilding and a couple of cultures are close enough to trade, invade or merge with eachother then radical changes will occur in your language too, maybe making some words irregular.
I'd assume analogy is also a useful tool in this area too?
Yes. "Dove" as a past tense of "dive" has no historical explanation except analogy to "drive".
It's always interparadigmatic analogy so if there's regular ablaut in a set of words or even just one common word people will use this for "regular" words to make them "irregular". Some linguists think this might be how the Semitic languages got their triconsonantal roots
There's another video about grammar by analogy. I think it may be on Ewa's channel, though.
@@senesterium Ewa?
@@DTux5249 Worldbuilding Notes is what Zeidra is referring to.
0:38 Why is Russian written in cyrillic but Japanese in romaji?
5:34 When did people say "I'm'na"?
4:16
The verb سأل in Arabic is perfectly regular though. أكل and أخذ too, since they follow a pattern specific to verbs beginning in glottal stop.
I was thinking back to your last episode realized that you be interesting to see a video combining the two concepts and seeing what a alien language you might sound like using speculative biology to see the kind of sounds other creatures might make and incorporate into a language
I needed a video from you right now xD. Thank you so much :)
5:07 Can you explain why intul was the perfect past tense of inta?
quick correction at 0:37
the verb "делать" means "to do/make" not "to say"
Thanks for the videos, such a great and very interesting one. Just remember that habeo is 'to have' in Latin, not in Spanish! That'd be 'haber'! Nontheless awesome vid
Raumanoetro has the adjective [go], meaning "equal" in the masculine singular. It becomes [jekwi] in the masculine plural, [kɛwa] in the feminine, and [kɛwɛ] in the masculine plural.
Edit: this was just the result of sound change.
pfap [p̪ap]
(child) proto-Bolen
pfapki [p̪apˈki]
(child-ᴘʟᴜ) proto-Bolen
pfap [p̪ap]
(child) SB
pfapki [p̪akʲ]
(child-ᴘʟᴜ) SB
In the orthography the word for child is regular, but in speech it isn't.
(I used what I call an "orthographical romanisation", where instead of writing the sounds you write a representation of an alphabet that doesn't have unicode, you write in a representation of it in the latin alphabet)
What about other grammatical irregularities that come from relics of the old language? English has one particular phrase construction that illustrates. "Here comes the bus." The "here" of that sentence isn't going anywhere. It is "the bus" that is the subject of the sentence even though it is in the object slot of that sentence. Similar phrases of "There goes ..." are also useful. Historically, "good" german or old english would place the verb second in the sentence, but as English lost its noun declensions for nominative and accusative cases, we gained fixed word order instead. But we retain this little phrase that make sense for us, but grammatically is odd. Do you have any recommendations for how to incorporate this kind of irregularity into a language and if there are other languages that have these little artifacts like English in them?
At 2:18 when there was word-final vowel loss, why didn't ”tundu” lose the final ”u” ?
Because word-final clusters are disallowed, so the "u" have to stay
4:17 all the irregular Hungarian verbs listed here have a double n in the infinitive, never really noticed that.
Not sure if this is a good idea or not, but what about just, NOT conjugating verbs, like, at all?
If I recall correctly, Mandarin does this, but I'm not 100% sure.
I mean, sort of. Morphologically, yeah, Mandarin has no verb conjugations, but can one really say that the different particles and modals around the verbs are not conjugations, AT ALL? I mean, there might not be any person or number agreement, but I would think of the various particles as some kind of conjugation.
@@Sovairu Sorry for the late reply.
But yeah, I was speaking morphologically, yeah. I'm sorry if I came off the wrong way.
I'm not sure if I'll include particles and such in mine or not, but I do know that I won't be modifying verbs in mine.
@@Yoshimaster96smwc Oh, no, your comment didn't come off the WRONG way; it just seemed a bit vague or incomplete. I mean, if you want the form of your verbs to never change, and have other separate words in the sentence do all the heavy lifting for person, number, TAM, voice, evidentiality, conjunction, or anything else, that's fine; go wild! I just think of it as sort of a gray area with saying that such a system does or does not have conjugations at all. However, to be fair to you, after I went back to the almighty Wiki to check on conjugations, it seems that they are pretty much understood to be different inflections on the verb, instead of separate words ( en.wikipedia.org/wiki/Grammatical_conjugation ). So, yeah, if you want your word for the verb eat to only ever look like "et" or something, then go ahead. It's perfectly fine!
0:38 there is a mistake. "делать" is not "to say", it is "to do"
all irregularity can ultimately be analysed as a very baroque "regular" system with lots of ifthenotherwises - somebody did this for Latin nouns. I forget his name, but I do recall that the nominative case gave him the most trouble.
In a few hundred years, I predict that the infinitive form of "can" (which English doesn't have right now) will be replaced by some ghastly contraction of "to-be-able-to", which will confuse language learners for centuries to come.
Jonty Levine probably “t’bable”
Tobat
TO-Be-Able-To
I think we should make defective verbs complete again. To can, to will, to shall, to may, to mote (the obsolete present form of _must)_ and all the others I forgot.
@@grillygrilly If you tell that to someone now, he probably cans not understand it.
@@GlaceonStudiosTrue.
If I wanted to reform an existing language instead of creating a new one, how would I do so? I’m looking for a way to make the English language more consistent, without actually changing any words or grammar, just by making a script that better fits the language. Any tips on where I should start? I’m thinking of taking a system like Hangul and transforming it into an Alpha-Syllabary thing, with the many irregularities of English in line.
English is a lost cause unless you want to write a script for one dialect of English. The number of vowels in English is different per dialect
Though irregularity is most often the wrong term. The "irregular" verbs of modern English still follow a regular rule. Only and are truly irregular as their paradigms have different roots which can't be predicted.
The past participle of go comes from another root, indeed.
More specifically, from wend, which means turn, return, and follows the same conjugation pattern as send.
Really, a regular rule predicts all of drink->drank, write->wrote, freeze->froze, wreak->wrought, make->made, split->split, sleep->slept! (and doesn't also predict think->thank, bite->bote, wheeze->whoze, sneak->snought, take->tade, sit->sit, and beep->bept!) I'd love to hear about this magical rule!
@@AzrgExplorers many of the irregulars you are listing can be predicted through ablaut relations, although English pronunciation and spelling can randomize it a bit. True irregulars would be verbs shifting to different roots, or taking suffixes other than -ed, unless there is a way to determine them through phonology.
@@alternateperson6600 If you have to study the etymology to even halfway predict the forms of verbs, what's the point in calling them "not truly irregular"? A language learner still has to just memorize them!
@@AzrgExplorers nobody has even said a thing about etymology, although some verbs which were once "irregular" (shifted to preterit through Germanic Umlaut) might have lost their ablaut inflection in the transition to Modern English. Sadly enough, it seems like no one has gone so far as to compile some sort of material with the ablaut relations for verbs which take in umlaut to change tense. It would greatly help English learners, hell, even native speakers, who think "irregular" verbs are completely a hit or miss dilemma. But because English is, somewhat, conservative on etymology, some ablaut relations haven't changed at all, even if the radical vowel has shown changes on pronunciation, so that Old English can serve as a valuable reference to compile such material from scratch.
Now... how do you make irregular verbs in a conlang which exclusively uses prefixes?
good question
Watch the whole video. Not all of them involve word final vowel loss
Hope this was helpful
There could be, like with i'mma, some form of contraction, or mid- word consonant loss.
Look at Navajo, for one, as it is exclusively prefixing (at least on verbs), and a lot of weird things happen when those prefixes start piling up. Or rather, do the same things as with any other affixes, sounds, or languages. This sounds flippant, but, really it's true. No matter what kind of affixes or morpheme combinations you have in a word, they will usually count as the one, whole word, and diachronic and synchronic sound changes will result based on the whole word.
About the same way. Instead of final vowel loss, you could have some syncope in pre-stress syllables. However, if it was me doing it, I'd introduce special sound changes just for intervocalic consonants, most likely some combination of progressive palatalization and vanilla lenition. If clusters are allowed, I'd probably also have some unique outcomes for certain clusters.
How to evolve ergativity?
Lol in Tamil (I'm a native) to be is completely regular! But we don't usually use a copula, so maybe it wasn't used in the old days as well.
Tamil definately isn't 100% regular
@@xeuxixiliak8417 I only said the verb 'to be' is regular
It's hard learning all of this in my native language, but it's disheartening to realise that it's even harder in English even when it's simplified to oblivion XD
One irregularity that really threw me off when I first saw it was the Breton word for "do" because its infinitive "ober" is a loanword from Latin "opera" but the conjugated forms ("graet" among others) are native verb forms similar to those in Welsh and Cornish. I know many languages, such as English with "be", combine different word roots for basic verbs but this the first time I've seen a language go as far as to borrow one of the roots used in their formation.
I have a question. I'm currently working on my first serious conlang, and I was wondering if it's naturalistic for a language to differentiate between alienable and inalienable adjectives and have them be in different positions in the sentence. For example "Taly teitrsy Fa" means, "I'm 16" and "16" comes after "Taly" (To be) because age can change easily. However, "Taly Fa batrda" means "I'm a man" and "A man" comes after the first person pronoun; "Fa" because gender is an innate thing.
My logic is that the "alienable" descriptions are explaining how the subject is IN THAT MOMENT and how easily changeable that state is. And the "inalienable" descriptions is how the subject HAS BEEN and how it is nearly impossible to change that description.
Thanks!
It seems logical enough to me, specially if your speakers' culture focuses on possession and things that are "of" something.
Yes, that occurs in some natural languages.
I'm particularly fond of the fossilized conjugations process. Like how in English alot of multisylabic words that end in -us become -i and and -a becomes -ae and both are pronounced "ai"
Those words were directly borrowed from Latin, along with their plurals. Hence, those plural endings.
Hearing that, it seems amazing that Japanese has so few irregularities especially verb-wise
I think that the English simple past tense used to be the imperfect but the its role as the imperfect was replaced by a past form of *to be* and the present participle. I believe that *to be* still has irregular past tense forms is due to it once being an imperfect.
I noticed that almost all of the irregular past tense verbs in English which replace the vowels consist of just one syllable.
That's an awesome video but I just want to point out that, out of all the Arabic verbs you listed, only ليس (not to be) is truly irregular in that it can only conjugate for past. All the other verbs in your list have "regular irregularity".
The verb "To be" in Swedish is:
To be: Att vara
I am: Jag är
You are: Du är
He/She/It is: Han/Hon/Den/Det är
We are: Vi är
You are: Ni är
They are: De är
So basically, it's always "är", although I have no idea how "vara" changed to "är".
Jinado The different forms are actually originally two different verbs which merged into one. är
Yeah, to be in Swedish is probably the most regular irregular verb ever.
I’m sorry if it is too much of a hassle, but could you provide the sources you read to understand the methods of grammar creation in your “how to” video series?
I’m in the process of making a conlang now but I’m having some trouble with Adpositions in particular.
Plus I’d like to be able to read the information myself for more reference.
Thanks for your help!
6:11 English "have" is not a descendant of Latin "habeo". I'm TriGgEreD!!!!!!
Could you provide the sources you use, please?
1:04 There's that "mouse" thing again. The easy way to "fix" it is to change "misi" to "mys".
Well, this is a bit weird. Ancient Greek too has this thing with good
Ἀγαθὸς (Agathos, good)
Βελτίων (Beltiōn, better)
Βέλτιστος (Beltistos, best)
The suffixes are nothing extraordinary but the adjective is considered irregular.
Also, for some reason, the verb to be in Ancient Greek is very regular, although from a rare category of verbs ending in -μι (mi), which is the oldest IE suffix, while normal verbs end in -ω (ō) even in modern Greek. This is surprisingly the most inexplicable suffix of Greek, where as the other person conjugations came directly from PIE. Apparently, it comes from Ἐγώ (Egō, meaning I) and it's made to comply with its suffix.
1:10 wtf Why is Irish among the romance languages?!?
It is believed that Latin and the Celtic languages had a common ancestor called 'Italo-Celtic'. Maybe that's why
Oh right, the tree makes sense after all, it was just strange to see it of the same colour...
Thanks
Now that I see that tree I'm lowkey attacked that Portuguese is not right there beside Spanish or Italian:') jk the video is amazing
2:08 fun fact, "flores" actually means flowers in Spanish.
And Portuguese, because of course it does.
no shit
So in a first proto-conlang I should do everything regular, right?
Does /l/ becomes /d/ when following all nasals?
I hope to see a video on the topic of mixed conlangs aka fictional creoles/pinjins, there is not much information on this topic in the conlanging community.
Oh, the copula. So irregular that Genki literally teaches it as if it weren't a verb, but some other class of word altogether (though still at the same time).
I read on Wikipedia that while the copula is a verb in most languages, it can be other parts of speech. In some languages it's a suffix.
My conlang has irregular verbs that only slightly deviate from the regular pattern (as in they are one or two missing or transposed letters away from being perfectly regular) and only in some of their conjugations (as in just two or three out of the six tenses are irregular). I'm not sure how realistic this is, and I'm wondering if I should try to kick the weirdness of these verbs up a notch.
This doesn't sound natural, thoguh again you didn't say much. Irregularities that stand out are more likely to remain than those that are a nudge away from being regular. Also, irregularities tend to be regular themselves, e.g. most irregular verbs will follow one of only a few subpatterns.
@@Релёкс84 There is a set of what I've dubbed "alternative verbs" that follow a different pattern and are regular amongst themselves. These are much more distinct from the common regulars, but I wasn't considering them truly irregular because there's so many of them and the pattern they follow is entirely predictable if you know which ones are alts.
is there anything about a natural language that can 'force' high regularity? the best i can think of is being a pidgin/creole, which i heard tend to be especially grammatically simple.
i could also imagine that speakers could enjoy the elegance or regularity, but maybe that'S wishful thinking. needless to say i'm more interested in engineered languages than naturalistic ones. :P
You can 'force' high regularity by using a language reform. Language reform can occur in two ways: simplification and purification. Simplification can reduce irregularities in a language while purification is another way to introduce even more irregularities.
The Academy, and planned languages (Ausbau, mostly).
That's basically when politics prevents the natural evolution.
That happens quite often, and even more for the written form (the best example I know of is castillan spanish that became almost phonemic with a single, intense reform).
If your language developed all its inflections recently, they're likely to be highly regular, maybe with a few predictable phonological rules like the "tundu" example.
@@senesterium yeah, language plannung seems to usually have a fairly auhoritarian nature. i'd like something more bottom-uppy, but maybe you'd need a society of linuists/conlangers for that..
How do you get multiple languages from one language then?
Different sound changes are the first step.
(they start of as dialects)
Example sound changes:
In Estonian vowels were dropped from the ends of words: keeli > keel
In Finnish long mid vowels became diphthongs: keeli > kieli
I present to you the forms of "to-be" in arthumonabuz:
thoý [θ̠oy] (imperfect)
oý [oy] (perfect)
dni [dni] (future)
the perfect "oý" used to be the verb for "to-rest" in the protolang but over time became the copulae, the imperfect: "thoý" uses the old past continuous, verbs in arthumonabuz usually don't have a future but to be is the exeption, "dni" is just an old verb for "to-grow"
Oh, you're doing what Hungarian does and only giving the copula a future tense then?
@@alexandernyberg8668 yeah :D
Got it. So I just make up a random irregular word, then _derive_ it from an older random word
I got an ad for Metamucil before this video.
I willed like to suggest that we all use regular inflections of verbs.
It be easiest for everyone if it's regular.