Language, Linguistics, & Grammar

[Catullus64]

Hopefully a thread for people who share my enthusiasm for these things, and want to ponder the workings of language. If not, then maybe someone can help me with a curiosity I've had for a long time.

The understanding I have about Mandarin (from both native and secondary speakers) is that it lacks a clear equivalent of the verb "to be." Aside from the fact that it must make translations of Hamlet very confusing, I've struggled to grasp what this means in practical terms.

In English, and all the languages I've studied (see below) the verb "to be" pulls a lot of syntactical weight.
1. Lots of verbs rely on it as an auxiliary verb to form certain tenses and aspects.
2. It's the primary means of predicating adjectives.
3. It has its own distinct meaning as an intransitive, like the example of Hamlet's soliloquy.

As a starter, could someone fluent in Mandarin, or other languages that lack a verb "to be", try to explain, in terms comprehensible to my very Indo-European ears, what words or syntaxes exist to fulfill those three functions at least?

I'm more generally interested in whether the unification of all of these functions in a single word has some universal basis, or even makes sense, but I think the discussion needs more percolation before that can be addressed.

Spoiler: My Language Background

Show

All of my study of languages has been in Indo-European languages; native English speaker, studied Latin in high school and Ancient Greek and French in college, picked up some Italian, German, Spanish, and more French for travel purposes.

[Metastachydium]

I'm not fluent in Mandarin (in fact, I don't speak it at all) but I know this and that about its grammar. Mandarin does have a copula and it also does have a verb that means 'to exist', but the two do not coincide. Further, the equivalent of a subject complement expressed by an adjective is simply the adjective in itself which „becomes a verb” (or, more precisely, functions as such).
As for auxiliaries, a language does not have to use them to construct any of its verb forms. Any tense or aspect can, in theory, be expressed through synthetic means. That said, Mandarin does have various particles that help in marking, say, aspect. These just do not happen to coincide with the copula.

[Xuc Xac]

I don't know much Mandarin, but I know some Cantonese and a lot of Vietnamese. The answer to your question is easy when you realize that the "be" in "to be or not to be" is not the same as in "I am eating". The first one is a verb that means "to exist" and the second one doesn't mean anything by itself and just indicates that the verb "eat" is in present tense (and progressive aspect shown by the "-ing".

For example, look at these two sentences:
1. "The dog is eating."
2. "Eating is necessary for life."

Even though they're spelled and pronounced the same, the words "is" and "eating" are different in those two sentences. In sentence 1, the verb is "eat"; "is" and "-ing" are particles that show tense and aspect of that verb. In sentence 2, "eating" is a gerund and the subject of the verb "is".

In Vietnamese, sentence 1 is "Con chó đang ăn". "Con chó" is equivalent to "the dog" (Vietnamese has no word for "the" and English has no word for "Con", but "Con chó" is used where English would say "the dog", but classifiers are a whole post in themselves, so don't worry about that for now.) The verb "ăn" means "eat". The word "đang" doesn't mean anything by itself. It's a particle that indicates that the following verb is in present tense and progressive aspect. Mandarin makes use of particles in the same way, but usually as suffixes to the verb or adjective.

In sentence 2, Vietnamese would use "là" for "is" for a copula (connecting two things together like "she is a doctor" means "she=doctor").

In most cases, Vietnamese doesn't even need a verb to predicate adjectives. (It uses it in sentence 2 because "eating" and "being necessary" are being compared like nouns rather than functioning like a verb or adjective.)

In English, we can say "the brown dog is eating" or "the dog is brown". "The brown dog" isn't considered a complete sentence in English even though saying "the dog is brown" provides no additional information. In Vietnamese, "the brown dog" and "the dog is brown" are both "con chó nâu", because adjectives function like stative verbs (verbs that describe a state of being instead of an action, like "thirsts", "hungers", or "exists").

This is what you mentioned as its own distinct meaning as an intransitive, like "to be or not to be". In Vietnamese, you can say "A is B" as "A là B" but just saying "A is" would be "A tồn tại" meaning "A exists". In translations of Shakespeare, Hamlet asks "tồn tại hay không tồn tại", "exist or don't exist". (It's interesting that they translated "or" as "hay" instead of "hoặc". "Hoặc" would mean the question expects one of two possible answers. "Hay" is the "or" you use for things like "do you want some coffee or tea?" where the answer might be "coffee" or "tea" or "yes" or "no". It's like the translator is leaving Hamlet open to the possibility that there might be another answer he hasn't considered.)

So, long story short: English (and other European languages) use "be" for a lot of different things, but that's not a necessary component of the word "be". Other languages serve those functions in different ways: conjugation, particles, using separate words for separate meanings. That last is probably the biggest one. Just because one word covers a lot of meanings in one language, that doesn't mean every language does the same thing. If you look at the spectrum of colors, English distinguishes between green and blue, but Vietnamese has one word ("xanh") for both. They can see the difference, but need to describe it with added words like "xanh as a tree leaf" for green or "xanh as the sky" for blue. Russian has two words for shades of blue that English has to describe with extra words like "dark blue" or "light blue".

[SirKazum]

Mandarin does have a copula and it also does have a verb that means 'to exist', but the two do not coincide.

Not quite, I'd say. (I'm a non-native Mandarin speaker, not really fluent but working toward it.) Copula (linking a subject with a predicate) can be done in different ways in Mandarin AFAIK. If you're just associating an adjective or something like that to a subject, it's a "zero copula" where you just put the adjective there with no verb. Example (I'm using just pinyin transliteration without tones for ease of understanding): wo hen hao, "I am well", where wo means "I", hen is a moderate intensifier "very" (which tends to be used by default in such copulas where there is no negation or other adverbs), and hao is "good". "I am not well" is wo bu hao (bu is a basic "no"). As said above, I've heard that the adjective kinda turns into a verb, which makes sense with how you can modify it by aspect or tense such as wo bu hao le "I have become unwell" (i.e. "I'm not well", but indicating a change in state). That may be getting beyond my grammatical chops though.

However... there is also a verb shi which can be roughly translated as "to be", and which in certain situations can be used to link a subject and a predicate, such as wo shi baxiren "I am Brazilian" (where baxiren = "Brazilian"). AFAIK it's mostly used (in its meaning as a copula) to link a subject with a noun or noun phrase, or with some specific referent such as ta shi gao de "he is the tall one". (Shi also has other uses, such as adding emphasis to a phrase.)

And to complicate matters even more, other verbs may be translated as "to be" in other specific situations... for example zai, which roughly means "to be (in a specific location)". For example, wo zai jia "I am home" (jia = home). (And of course, zai also has a bunch of other uses, such as indicating continuous action, because why not.) I can't think of any other examples right now, but the way Chinese is, I'm sure other verbs are eventually going to mean "to be" in specific contexts.

[Aedilred]

A linguistics thread, awesome.

I don't speak Mandarin, but I do have a general enthusiasm for linguistics.

1. I think that even among Indo-European languages, English is not necessarily typical in leaning so heavily on "to be" for the imperfective aspect. As you will know from French, the imperfect is formed without an auxiliary at all, and in the present tense, perfective and imperfective aspect are generally distinguished by context rather than morphologically. I don't think there's any reason why you couldn't say "je suis dansant" (and indeed I'm sure many English-speaking students of French do) but it's hardly idiomatic. "Je suis en train de danser" is closer to the mark, but that's introducing some additional baggage and can't really be considered the same use.

Among other Romance languages, Catalan, Portuguese, and (as you know) Spanish, all use "to be + gerund" to mark present continuous, in a similar way to English, although not necessarily so frequently as English does.

Italian also marks the present continuous with auxiliary + participle, but uses "stare" rather than "essere". (Note that Italian does use "essere", together with "avere", for formation of the perfect, in the same way as French uses "etre"/"avoir".) "Stare" could probably be translated as "to be", but has a more locative sense than an existential one, and the principal reason to translate it as "to be" is precisely because of its use for the present continuous, rather than because as an isolated verb it has the same meaning as the English "to be". Again, Italian uses this form less frequently than English in any case.

My Romanian is terrible, but I don't think it marks for present continuous at all.

My knowledge of Germanic languages is nowhere near as good as of Romance, but it seems (from first-hand study) that the Scandinavian languages don't do it, and (based on Comrie at least) German doesn't really do it either, or at least not very much. I gather (Comrie again) that Dutch does it a bit, but not to the same extent as English.

Slavic languages are even further outside my comfort zone, as it's been years since I studied Russian and I was never particularly good at it, but I think Russian marks continuous forms with direct inflection of the verb, rather than by use of an auxiliary, whether "to be" or otherwise. From what I recall of Greek (an equally long time ago, and I an equally poor student) it does the same.

There is another western IE language I know of which relies even more on "to be" for forming verbs than English, and that's Welsh, which I think is representative of Celtic languages in general in that respect. Indeed there "to be + participle" is the standard way of forming verbs in every tense. If I had to guess as to where English got it from, that would be one of the prime suspects.


2. This is something which is definitely normative in IE languages, but not necessarily elsewhere. Even in English pidgin (and certain forms of internet-speak) the copula can happily be omitted and while the resulting phrase may not be strictly grammatical, it's perfectly comprehensible ("he thicc"). I gather that there are some languages which effectively dispense with adjectives, at least in that use, altogether by shifting that into a verbal expression. i.e. for a statement such as "he is brave", instead of having separate words for "he", "to be" and "brave", and saying "[he] [is] [brave]" you would have a verb "to be brave" and use the form "[he] [is brave]". I can't remember which languages do this, but I think the Austronesian family has some examples.

This is, incidentally, also a feature of my conlang, Sūwenanda.

I gather that there is a school of thought that adjectives as we think of them don't really exist, and that what we think of as adjectives are either a peculiar noun form or verb form, depending on use. I can see where it's coming from, but I'm not convinced.

3. Unlike the previous two instances, I don't think there's any morphological workaround where a language simply lacks the verb. It seems, based on the responses you've had so far, that usually where it appears to lack a direct translation, it's not because the languaged doesn't have a means of expressing it but rather because they don't have a single catch-all verb in the way English does, and instead various more specific words are used. I'm not aware of any languages that lack a word for "to be" altogether but if any exist, then it would indeed be hard to translate Hamlet's soliloquy. Of course, translating anything directly and exactly between almost any languages is very difficult anyway, thanks to idiom.

[SirKazum]

Among other Romance languages, Catalan, Portuguese, and (as you know) Spanish, all use "to be + gerund" to mark present continuous, in a similar way to English, although not necessarily so frequently as English does.

Italian also marks the present continuous with auxiliary + participle, but uses "stare" rather than "essere".

Just a note though, in at least Portuguese and Spanish (I don't remember how it's like in Catalan), present continuous is also done with the auxiliary "estar", which is cognate with "stare".

Also, as a curiosity since we're here, Portuguese is kind of an oddball among Western European languages (Romance languages + English) in that our past perfect is exclusively inflectional, not done with an auxiliary. The construction "haver"/"ter" + past participle, which is equivalent to how other related languages do it, has a completely different meaning in Portuguese (roughly equivalent to "have been doing").

2. This is something which is definitely normative in IE languages, but not necessarily elsewhere. Even in English pidgin (and certain forms of internet-speak) the copula can happily be omitted and while the resulting phrase may not be strictly grammatical, it's perfectly comprehensible ("he thicc").

AFAIK, African-American Vernacular English (AAVE) uses a zero copula in most situations (e.g. "he good"), as well as no auxiliary in some uses of present continuous (e.g. "he doing"). And, differently from standard English, it actually has two distinct present continuous tenses, both without and with the auxiliary (e.g. "he be doing"), which have distinct meanings and allow for nuances that don't exist in standard English. I don't really know that much about AAVE, but I think it's a fascinating subject, since (from what I've read) it's a distinct dialect with a whole well-defined set of grammatical rules that are distinct from standard English, which most people tend to discount as just "bad/incorrect English" since it's a low-prestige dialect. Buuut that's veering into a whole other topic that's not for this thread and probably not for this forum either, so I'll leave it at that.

[Aedilred]

Just a note though, in at least Portuguese and Spanish (I don't remember how it's like in Catalan), present continuous is also done with the auxiliary "estar", which is cognate with "stare".

That's a really good point which completely slipped my mind. Catalan also uses "estar". So in fact the typical Romance construction is to use a different word for the intransitive "I am" versus the present continuous construction "I am [doing]".

French être is also presumably cognate with "estar" but it seems to use it to cover the meanings of both "ser" and "estar" ("essere/stare" in Italian) in addition to the present continuous being largely absent.

Whether these are features that French has lost versus its cousins, or an innovation from a southern-Romance sprachbund I'm not sure. The former seems more likely except that, from what I recall, Latin didn't distinguish aspect in the present tense, or at least not commonly.* It has probably been researched; I'll have to look it up.

*This brings me back to my suspicion about a Celtic borrowing. All the western dialects of Vulgar Latin except Italian would have at least a partially Celtic substrate: I can imagine an innovation in Proto-Italo-Romance on that basis which survived in Iberia and Italy but was lost in northern France under Germanic influence. But then again, it seems Sardinian has it, and as far as I can recall that's considered the most archaic of the Romance languages so that wouldn't necessarily account for its absence in the eastern Romance languages.

AFAIK, African-American Vernacular English (AAVE) uses a zero copula in most situations (e.g. "he good"), as well as no auxiliary in some uses of present continuous (e.g. "he doing"). And, differently from standard English, it actually has two distinct present continuous tenses, both without and with the auxiliary (e.g. "he be doing"), which have distinct meanings and allow for nuances that don't exist in standard English. I don't really know that much about AAVE, but I think it's a fascinating subject, since (from what I've read) it's a distinct dialect with a whole well-defined set of grammatical rules that are distinct from standard English, which most people tend to discount as just "bad/incorrect English" since it's a low-prestige dialect. Buuut that's veering into a whole other topic that's not for this thread and probably not for this forum either, so I'll leave it at that.

I agree. I think there has been a tendency to disregard all lower-prestige dialects as being bad/incorrect until relatively recently even in serious circles, and many lower-prestige dialects still have yet to gain that recognition outside the linguistic-study sphere.

In England, the array of regional dialects is now taken as a sign of diversity and regional pride, with some of them showcasing archaic features which are now subjects of interest instead of derision. But even then, mockery of such is far from extinct in informal settings, and an ongoing shift from derision to validation has been observable even within my lifetime. And those dialects are often conservative ones where users can cloak themselves in a flag of tradition: newer and more innovatory dialects like Multicultural London English or AAVE don't tend to be given the same level of serious recognition.

[Corian]

If you want an interesting variant, Tibetan also has two axes of distinctions, hence 6 forms for "to be".
First, the frequent distinction between equational copula (the table is blue) vs existential/locational (the table exists in that room)
Second, the "testimonial" distinction: Whether this is personal sense experience, hearsay, or common knowledge. (I was only introduced to personal vs hearsay when I attempted to study Tibetan, the common knowledge was folded into one of the other two, but it seems that was a dielectal variant anyway.)

[Catullus64]

Alright word-heads, here's another quandry to contemplate. This one's a little more thought-experimental than the last one. What do you think is the origin of grammatical gender?

The thing that always makes me think about this is the example of Latin. There are several nouns that are inflected identically to feminine nouns of the 1st Declension, but which are treated as masculine; pronouns and adjectives which use them as referents are inflected masculine. Most of these are the names of traditionally masculine professions (nauta, sailor, poeta, poet, etc). French has a similar phenomenon with nouns that are gendered differently based on whether they are singular or plural.

This phenomenon, for me, calls into question the notion that grammatical gender originated as an expression of natural gender, or gendered objects like animals and people. (I'm taking it for granted that Latin, and other languages which I know to share this phenomenon, also share this quirk with other languages that feature grammatical gender.) If grammatical gender was developed to express a strictly gendered view of things (even non-gendered things, like tables and houses), why are poets and sailors not simply inflected that way?

What I theorize, therefore, is that grammatical gender originally had some function of classification not strictly related to natural gender at all; the separation of naturally male, female, and neuter subjects along these lines was mostly incidental, and only in later times, when society was more developed and gender roles more entrenched, did the speakers of the languages come to regard these linguistic categories as masculine or feminine per se. I take as support for this theory the fact that some languages (like, according to Wikipedia links, Kurdish) have classifications that can only be rightly called grammatical gender which have no correspondence to natural gender.

I have a hard time guessing at what such a hypothetical original classification could have been, however, and I'm interested to see if anyone can draw their own knowledge of languages into the thought-experiment. If you're a primitive language-speaker, the imaginary person who decided to start inflecting things differently along broad lines, what would your motive have been for doing so? What line do you draw?

[SirKazum]

The thing is, like a lot of linguistics discussions, it's easy for us to fall into the trap of thinking that Indo-European languages (or even just European languages) represent all there is to say about language, when there is a lot more diversity than that in the world. Gender is one such example; this idea of gender as being masculine and feminine (and possibly also neuter, depending on the language) is an Indo-European thing. Other languages may have completely different conceptions of grammatical gender (for instance, Swahili has 16 grammatical genders, but they don't include any male/female distinction), or no grammatical gender at all (for example, there's nothing in Chinese that corresponds to grammatical gender; there are different male and female versions of the 3rd person pronoun ta, i.e. "he/she", but that's a recent "Westernizing" innovation).

One theory I've heard is that grammatical gender is a redundancy feature of language (one of many, in fact), which helps disambiguate things. For example, if you're talking about a man and a woman, in the phrase "he looked at her" it's pretty clear who's looking and who's being looked at. Without a gender distinction, say for example the third-person pronoun was X, it would be "X looked at X" which would be ambiguous. Of course, there's still plenty of room for ambiguity when you're talking about more than one referent of the same gender (e.g. "he looked at him"), but hey, sometimes it helps. Under this light, gender doesn't necessarily have to do with "male vs. female"; while that's a handy distinction, which explains why it's often used, it's not the only practical one.

[SirKazum]

Chinese does use classifiers, which are similar to "gender" in the sense used in Swahili. It's hard to compare that with grammatical gender in Indo-European languages because we discuss grammatical gender in terms of how it affects inflection, but Chinese basically doesn't have inflection of any kind, so the Chinese noun classes obviously don't affect inflection the way Indo-European genders do.

Fair enough. I guess I was looking at it through the lens of grammatical inflection, but Chinese classifier categories seem to be more or less the same thing as Swahili noun classes (I don't really know much about Swahili), only without the inflection aspect as Chinese doesn't really have any of that.

[DavidSh]

I sometimes see the Indo-European grammatical gender as analogous to the vowel harmony systems that show up in languages such as Turkish or Hungarian.

[Catullus64]

I sometimes see the Indo-European grammatical gender as analogous to the vowel harmony systems that show up in languages such as Turkish or Hungarian.

Talk more.

[DavidSh]

Talk more.

But then I look more closely, and things fade away.

Once, in preparation to a trip to Turkey, in connection with a solar eclipse, I tried to teach myself some Turkish. It's mostly gone now, but if I recall correctly, they divide vowels into a number of classes, and words with stems with vowels in one class get inflectional endings with vowels in the same class. So I imagine proto-Indo-European might have been assigning inflectional endings based on what sounds they thought went together, rather than any thought of assigning sexes to things that aren't animate.

(I've heard that Hungarian has its own form of vowel harmony, but I've never had occasion to learn Hungarian. Despite the excellence of Hungarian mathematicians.)

[Vinyadan]

The thing is that languages all build upon what they previously had, and we can only go so far in the past.

So we are left wondering why and how certain things came up. I look at Catullus's post, it's not that I'm saying that he is wrong-wrong, but the issues he raises refer to historical grammar, and cannot be answered simply by going with Latin, first declension. So you have to look at language history, and you start going back to proto-Italic and proto-Indoeuropean. And it is important that both questions and answers refer to precise languages. Once you do that, you can do comparisons to other languages.

So you have to say, how did the first declension even come to be? Is it really a feminine declension? Latin doesn't really have any such feminine declension. (Yes, adjectives and pronouns with feminine in -a; but here it's about nouns). Of the five declensions of Latin, three have all three genders, and two do not have neutrum; and neutrum is the only one to have a different nominative and plural accusative.

The Latin first declension is the descendent of the proto-Indoeuropean first declension, although it contains some innovations that is shares with the Latin second declension, some of which are drawn from the pronominal declension.

The proto-Indoeuropean first declension was overwhelmingly, maybe even exclusively made up of feminine names. The nominative singular was identical to the plural of certain nouns, because these plurals weren't really plurals, as much as a singular collective noun (like cow -> herd). However, Latin has some masculines in the 1st declension that aren't loanwords. Agricola, scriba, parricida. What are these? They all contain a verb and express an action (field-worker, writer, father-killer).

And we also have a close relative to the Latin first declension in the Greek first declension. Here male names are frequent, and they also get an indicator feminine names don't get, an -s (a common nominative indicator). Nauta e.g. is the Latin adaptation of Greek nautas (ναύτας).

Fun fact: we also know that, in Old Latin, parricidas was also written with an -s. And there are more examples of this sort. So is this masculine -s in the first declension a new thing that sometimes popped up by analogy, or was it already in proto-Indoeuropean? We don't know.

By the way, many Greek 1st declension masculine names in -s are clearly derivatives: nautas from naus (ship), poietes (Lat. poeta) from poieo = I create, stratiotes from stratos (soldier and army), hoplites from hoplon (arm or tool), polites (citizen) from polis. (the -as vs -es endings are both 1st declension). I didn't look them all up, but it seems clear to me that they all are formed by joining a root with a suffix that takes on the first declension, and maybe the Latin masculines of the first declension have a similar history.

And, to tell the truth, it's uncertain (but very likely) that proto-Indoeuropean had feminine nouns. The oldest of its descendant languages (e.g. Hittite) only have animate vs neutrum. It could be that such languages simply removed the masculine-feminine distinction (for example, language change could have made them indistinguishable), or that they stayed true to the original binary proto-Indoeuropean system and all the other languages instead innovated from this point of view.
To add to this, Greek has some adjectives that don't have masculine, feminine and neutrum, but only two declensions -- neutrum and the rest (-os and -on).

Now, about the distribution of masculine, feminine, and neutrum in proto-Indoeuropean. Neutrum was used for inanimate things. The animate cases (masculine and feminine) were used for animate creatures. However, there probably were alternative ways to refer to an inanimate thing through an animate gender to express that it was an agent (like fire, "I lit the fire" but "the fire burns the tree"). Why active fire would become masculine and active water would become feminine, that's a different problem, the one at the heart of Catullus's question, and I don't know what the answer is. However, I am fairly certain (from comparison from memory of the daughter languages; the now too old to be reliable dictionary by Pokorny doesn't have m-f-n tags) that people nouns follow the gender you would expect (father, mother, sister, and so on).

EDIT: Just one thing, the thing with vowel harmony: phonetics definitely have an influx on grammar, although it can be mitigated by other elements. For example, Italian doesn't have neuter anymore. Why? Two reasons: the first is that Italian words were built from the accusative. The second is a phonological development: consonants at word's end got cut off. So e.g. all the accusatives dominum (M), manum (F), and corpus (N) lose their final consonant and get a -u ending. However, when you add a (proto) article, you get illu dominu, illu corpu, and illa manu, so there is no difference any more between masculine and neuter, but there still is a difference between masculine and feminine. Result, la mano is still an irregular feminine today, but il corpo is masculine.

[ForzaFiori]

Hungarian does have vowel harmony, it's a huge part of the language. There are fourteen vowels - a, e, i, o, u, o and u with umlauts, and then all those vowels accented (the umlauted ones have a double accent). those based on e, or the umlauted o and u are "front" vowels (the sound is made in the front of your mouth, similar to "slender" consonants in Irish) those based on a, o, or u are "back" vowels (made in the back of your mouth, like "broad" consonants in Irish), and those based on i can be either one, you just have to learn each word individually.

The vowel harmony is important because Hungarian is agglutinative, meaning it is built with LOTS of suffixes. Those suffixes have different spellings based on if the last vowel in the stem is front or back. As an example, "ház" is Hungarian for "house," and "epulet" is "building." The suffix "ban" or "ben" means "inside." "Inside the house" is "a hazbAn" ("a" is "the" - very confusing for an English speaker like myself lol) but "inside the building" is "az epuleEn." (the z in az is like the n in an - used when the next word starts with a vowel).

Some suffixes have a third ending for when the final vowel is the umlauted o or u or the accented versions, but not all of them. Usually those vowels are just considered front vowels.

Also, kinda similar to Mandarin not having "to be," there is no verb for "to have" in Hungarian. Instead, you would say something like "My dog is" to mean "I have a dog" and "My dog isn't" to mean "I don't have a dog." Definitely takes some getting used to. It's also one of the few non Indo-European languages in Europe, and is completely unrelated from any of it's neighboring languages.

One last fun fact about Hungarian. All of the vowels with accents that aren't umlauts were originally written as a doubled vowel - so "aa" instead of an accented a, which makes Hungarian begin to look a little more like it's distant relative, Finnish.

edited because the forum does NOT like the way Hungarian vowels are accented apparently.