1. The WordReference Forums have moved to new forum software. (Details)

Separate words for he/she, but no grammatical gender

Discussion in 'All Languages' started by john_riemann_soong, Jun 12, 2007.

  1. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    Is English the only one that does this? I note that in most languages that distinguish between "he" or "she" (Chinese for example, uses one pronoun for the third person singular) they usually have grammatical gender too. Most languages, it seems, that do not have grammatical gender (e.g. Chinese and Malay) will not distinguish between "he did it" or "she did it either".

    English lost grammatical gender fairly recently. Can one assume that looking at the trend of history it's only a matter of time before we use one word for both "he" and "she"?

    Some interesting cases I suppose are the Japanese honorifics system (which I assume declines for the gender of audience and speaker) and Vietnamese which seems to use third-person for self-reference a lot.
     
  2. Outsider Senior Member

    Portuguese (Portugal)
    I believe the same is true of several Germanic languages.
     
  3. perfavore Senior Member

    USA
    Philippines - Tagalog
    Have we stopped using he, she, his, her in English?
     
  4. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    No, but it seems that most other languages that drop grammatical gender eventually do, if the predecessors of Chinese and Malay (I assume for other Austronesian languages like Hawaiian or Tagalog it's the same thing ) once had gender, but eventually dropped them.

    I suppoes that could be a faulty assumption too, but I note that the older languages tend to have more noun classes and more inflections, so I'm guessing that many non-gender languages lost gender distinction, rather than the current gender languages gaining gender.

    That is not grammatical gender. Grammatical gender is making declension based on the gender of a noun.

    For example, in French:

    un bon fils
    une bonne fille


    Here the root fil-, signifying descendants (from where we get filial), is declined into a feminine -ille and a masculine -ils ending (from L. filius and filia). The articles are declined into masculine and feminine (un and une) and adjectives have to agree with the noun's gender (bon and bonne).

    Old English had to do the same thing too (though it had weak and strong declension, which isn't a feature of the Romance languages). I don't know enough OE to give you an example off the bat, but suppose that we had grammatical gender in English today, and suppose that in this day and age the masculine form of "good" is "godo" and the feminine form is "goda" (let's just say), and that the masculine form of "child" for the nominative is "child" versus feminine "childe" (two syllables). So in my newly invented modern English with grammatical gender, you would have to use different forms of "good" to say "good boy!" versus "good girl!" (goda childe versus godo child).

    (Students of OE will probably be horrified that this does not reflect at all the pattern in which OE nouns and adjectives were declined but I just needed an example. :p)

    Ah, I thought all the other Germanic languages had stronger inflections than English did.
     
  5. French certainly has grammatical gender

    he dit it : Il l'a fait
    she did it: Elle l'a fait

    As does English, where, as far as I know, there remains a clear gender distinction... And a mix-up is noticeable and causes confusion.

    The only language I personally know of that never distinguishes, even in the personal pronouns (I, you, he, she, we, they) is Turkish. There is no difference whatsoever. But in English, when you consider the personal posessive (his car, her house), it is absolutely pertinent.

    Maybe you could rephrase your question? I think you probably have a very good question, but I have missed the point. :)
     
  6. modus.irrealis Senior Member

    Toronto
    English, Canada
    For Germanic languages, there's Afrikaans, which Wikipedia says has no gender but gives the personal pronouns hy-sy-dit for he-she-it. I'm not sure if any of the other Germanic languages have lost grammatical gender.

    Could you say written Chinese counts, since there's three different characters for the 3rd person pronoun reflecting gender?
     
  7. Ah, just saw your post, and I'd like to add that although in French you have sa fille, where the agreement is with the gender of the noun, in English you have her daughter, where the agreement is with the second person pronoun, instead of the noun. It's a different system, but still, you cannot forget it's there and say "his car" when you mean "her car".
     
  8. Well, then, for sure there's Turkish, which makes no distinction between he-she-it: o-o-o. Third person they-it: onu-onu-onu. And the posessive also does not change, neither according to the gender of the subject nor to the noun. It is a gender-free language. But how could English be gender-free aside from attempts at political correctness?
     
  9. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    Mmm, this is because the dialect went through a sort of creolisation process in the mixing of cultures in South Africa, I suppose? (Sad that this however, did not stop apartheid ...) I don't know if we could call the intermingling between Norman French and Old English "creolisation" but the process does seem sort of similar -- you tend to develop analytic languages that drop inflection. Perhaps the ancestors of Chinese (ultimately, the ancestors of Sino-Tibetan) went through a similar process once. I assume Tibetan uses one pronoun for both "he" and "she" too.

    I don't know the other Chinese dialects but I have a thought -- the different writing pronouns sound the same in Mandarin but perhaps they do not for the other dialects. Mandarin as you know has a tendency to merge ending consonants.

    But I checked the etymology for all three characters and apparently in Old Chinese (which is pretty old indeed! Sinologists suspect Old Chinese didn't have tone), they would all have been pronounced as "slaj".

    Something reflected in pronunciation as opposed to just writing to me is more reflective of concepts native speakers have in mind. I am in a French gaming clan where the -t ending is interchangeable with -s ("il me suis!"), and "toutes" is regularly interchanged with "toute" -- because the distinction between these endings is not reflected in pronunciation and they are lazy and often they type in chatspeak. ;)

    But you notice, they never mix up "cettes" and "ces", or "bon" or "bonne" or "oeil" versus "yeux" etc. -- they only have the concepts (as native speakers) that you can perceive in pronunciation. It would seem something that would be only distinguished in writing would be artificial. I think it is because that I am a non-native speaker that I am one of the rare adolescents in the clan who actually observe the distinction between "veux" and "veut". ;)

    Anyhow, if you consider the fact that before the modern era, most peasants in China were illiterate, such a gender distinction only in writing doesn't seem to be a concept that would be naturally "felt".
     
  10. Outsider Senior Member

    Portuguese (Portugal)
  11. perfavore Senior Member

    USA
    Philippines - Tagalog
    I was born without the distinctions and we never studied if there were any.
     
  12. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    Well, the pronouns I suppose are the most resilient parts of a language, because they are so commonly used. Nearly all IE languages have their second person singular pronoun not straying very far from "tu". (Hence, why bhrata in Sanskrit and mater in Latin sound so similar to brother and mother). Hence, I suppose the last parts of English declension are basically clinging on in those pronouns.

    But I wonder if the distinctions too, will eventually disappear. Grammatical gender is not a productive concept in English, just like we no longer use the plural -ren ending productively (though we have children and brethren). We for example, do not make distinctions between accusative and nominative cases save for our pronouns -- we do not say, "I have a computer" versus "I sent the file to your computerum" versus "I want to hit this computern!" (It is probably because of the fact that the sudden flood of Romance words into English were so hard to accomodate all at once into the declension of English that many productive forms of it were eventually dropped in Middle English.)

    Technically, we have very limited gender declension because we make a distinction between "his car" and "her car", but for example, we make no agreement with adjectives at all. The declension is so limited is nearly negligible -- English effectively doesn't have grammatical gender, just like English is effectively analytic even though it technically has a limited case system.

    For example, we do not make a real semantic distinction between "computerum sent I" and "computer sent me" (I am using "compterum" as a sort of stand-in for a modern English dative case, if there were one ;)). However, in a language with a productive concept of case declension, the first would mean "I sent it to the computer" versus "the computer sent me [somewhere]". Similarly productive gender distinction doesn't exist; in French you can keep track of two different "it"s if one is masculine and the other feminine (il la prend) -- in English "it takes it" would be nigh meaningless.

    The limited case system I suppose is sufficient to allow for easy understanding of declension when learning other languages, since it provides a point of reference. I remember in preschool in Singapore where we were going over how we do not say "that is him book" or "I saw they, and I wondered why "her" was so inequal to "him" -- it only has two case distinctions instead of three for the masculine (him, his, he). I suppose I must have used the concept of pronoun case agreement unconsciously before but before my teacher pointed it out to me I didn't consciously realise what I was doing.
     
  13. modus.irrealis Senior Member

    Toronto
    English, Canada
    I think that's what john_riemann_soong is asking, whether languages without grammatical gender are like Turkish in that they don't have sex-based distinctions in their pronouns, or like English which does.

    But about English not having grammatical gender, for me it's that English functions so differently from languages that do have gender. The choice between he, she, and it depends on something very different than the choice between il and elle, and English has nothing like the sort of adjective agreement that many languages with gender do. That's why I'd say English has no gender. I mean, I don't think the simple fact of having words that specify gender is important, because then I would think all languages would have gender -- Turkish has words like kadın, kız, aygır, kısrak, e.g.

    Although, if creolization did take place and the tendency is towards simplifying, maybe the fact that the different third person pronouns survived (so far) suggests that there isn't any sort of push towards eliminating different third person pronouns even when grammatical gender has been lost. Although, I guess with both English and Afikaans, the timelines involved aren't that great as of yet.

    I have to say that word doesn't look so Chinese :D. But is the "j" there the pinyin "j" or like the "y" in "yes"?

    I agree, especially since you're interested in seeing if this predicts something about the future of English. It seems reasonable that spoken languages are more likely to give clues about how languages develop, both because we're all native speakers rather than writers and written languages will always have some artificiality to them.
     
  14. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    I never was that piqued before, but now having checked the etymology of the different pronouns that all sound the same I am duly fascinated. Chinese writing developed from Shang oracle bones -- if they wrote different characters for different genders yet pronounced them all the same I would be truly amazed.

    What I am suspecting though is that scholars developed the different characters, all with the same pronunciation, which found its way from Classical Chinese into vernacular writing. I am not sure -- I am not an expert on Chinese character history and am sorely in need of someone who is. But for example, if the specific gender characters are found to be much younger than the one that covers both he, she and neuter, then that would be a giveaway that the characters don't reflect an actual natural concept of Chinese.

    If however, the characters are all of the same age, then that would truly pique me. Common characters like these are easy to find the etymologies for,
    Pinyin "j" IIRC represents a phoneme only found in Mandarin -- IIRC it is merger of "palatalised k" (/k^j/, like if you pronounced "kiosk" as one quick syllable) and some other phoneme I forgot. Hence, Peking => Beijing. ;)

    I only got interested in historical Chiinese phonology recently, hence I don't know all the written representations, since I got that off an online reconstruction resource developed by a Russian sinologist. I think it's "j" as in /aj/, like how white is /wajt/ -- a dipthong, effectively, since if my meory serves me correctly I have looked up words that give endings like "-aw", which I assume is /aw/ as in "house" (/haws/).

    And I know it looks radically different! Old Chinese is fascinating, isn't it? It's fun imagining that all those Taiwanese historical dramas in Mandarin about Qin Shihuang or Liu Bei or the Song-Liu conflict would have a significantly different atmosphere to them if they were to actually represent the spoken language at the time for historical accuracy. (I suppose it would also be fun imagining Robin Hood speaking in Middle English ... )
     
  15. jonquiliser

    jonquiliser Senior Member

    Headquarters
    Svediż tal-Finlandja
    This is true also of Swedish (and perhaps also Norwegian); there is no masc-fem. gender distinction, but there is still a grammatical distinction of common gender (words that formerly had either masc. or fem. gender) and neutral words. And there are gender-specific pronouns, "hon" and "han".
     
  16. Joannes Senior Member

    Antwerp
    Belgian Dutch
    Count (Standard) Dutch in: in declension there's only a distinction between gendered and ungendered words. But there's a triple gender distinction for the 3SG pronoun: hij 'he' / she 'zij' / het 'it'.

    In most (substandard) Belgian Dutch varieties, there's still a difference in declension between masculine and feminine nouns, however. The subject was recently touched upon here (in Dutch).
     
  17. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    Distinguishing common gendered nouns and neuter is interesting; I wonder how it is taught? I mean, usually we see neuter being dropped (like from L. to Romance) or the whole thing being dropped (like in English), but I suppose if the gendered forms were closer to each other than it was neuter, it would be just a phonological process.

    Are there non-Germanic languages which have no grammatical gender but distinguish gender pronouns?
     
  18. modus.irrealis Senior Member

    Toronto
    English, Canada
    According to Wikipedia, it looks like these gender-specific characters are fairly new and due to influence from European languages. There even seems to be a God pronoun 祂 "He", so this does seem pretty artificial.

    Thanks.

    That's true. With languages that have historical spelling systems (like English does to a great extent) it's easy to project modern pronunciation into the past and not think that much has changed. But I imagine reconstructing Old Chinese pronunciation must be really difficult because of the script.

    But about the actual topic of the thread, going through Wikipedia I found two other languages:

    Basque seems to have no gender but distinguishes masculine and feminine in the 2nd person singular with verbs (I don't understand the details though),

    And Thai also has no gender but distinguishes masculine and feminine in the 1st person singular.

    So I guess it's possible that the sex-based distinctions can occur with other pronouns beside the 3rd person as well.
     
  19. Whodunit

    Whodunit Senior Member

    กรุงเทพมหานคร
    Deutschland ~ Deutsch/Sächsisch
    It doesn't work for German, since it has theee genders (masculine, feminine, neuter), and this has always been the case (in MHG, OHG, and - as far as I know - the East Germanic and ancient Norse Germanic languages, there were three genders, too).

    The three German 3rd person singular pronouns are er, sie, es. The interesting thing is this:

    Dort steht ein Tisch. Er muss geputzt werden.
    There is a table. It has to be cleaned. (in German, it is "he" instead of "it")

    Ich möchte meine CD wieder haben. Kannst du sie mir geben?
    I want to get my CD back. Can you give it to me? (in German, it is "she" instead of "it")

    Das Glas ist heruntergefallen. Heb es bitte wieder auf.
    The glass fell down. Please pick it up. (both English and German use "it")

    In German, we decide whether the noun the pronoun refers to is of masculine, feminine, or neuter gender. Persons, objects, and abstract things are treated all the same. You can still see things like this in English for ships and trains that are sometimes referred to as "she", as far as I know.

    I would say it as follows (just in case you want to know it :p):

    se goda geongling (good youngster)
    sêo gode gyrele (good boy)

    I should add that gyrele a) is a reconstructed form and is not attested, as far as I know, and b) was used for either sex. However, you should know that (the otehr way round as in German today, if I understand the OE grammar correctly), there are two kinds of adjective declension: weak and strong. The weak declension is the one that looks distinctive in the masculine and feminine nominitive, so I have chosen it with the determiner (the weak forms have to be used if they depend on a determiner).

    See here.

    Did the consonant combination of s and l really work in Ancient Chinese? The word doesn't look quite Chinese to me either, but you're the native speaker and I believe you. :)

    Nevertheless, I wonder how Sinologists could have reconstructed some pronunciation for Ancient Chinese. The characters may have changed and it is possible to reconstruct written things, but how can you tell from characters how their were pronounced then? I could imagine the comparative method of modern pronunciation, but Ancient Chinese would be too old for that. Did they (Ancient Chinese people) maybe used something like the Japanese furigana?

    The same happens in English with no and know, wear and where, ...

    They can't, because cettes doesn't exist. :p

    In all (?) Indo-European languages that use a case system, the verbs to have and to hit both require the accusative, so it should be as follows:

    singular:
    nom.: This is a computer.
    gen. The computeres program does not work.
    dat.: I sent the filen to your computerum.
    acc. I have a computern.

    plural:
    nom.: These are some computers.
    gen.: The computereses programs do not work.
    dat.: I sent all the filenes to your computerumes.
    acc.: I have three computernes.

    Again, just invented, but it's more consistent than yours. :)

    You can't do that without any cases. Compare these:

    German: Der Hund aß den Fisch. AND Den Fisch aß der Hund.

    I would tell you (if you were a beginner of German) that Hund means dog, is the past tense of essen, which means to eat, and Fisch means fish.

    Do both sentences mean the same to you after all? :)
     
  20. Qcumber Senior Member

    UK English
    What you say is funny because in linguistics we use personal pronouns to test whether a language has grammatical gender or not.
    The fact that English has three pronouns for the third person singular proves it has three genders: masculine, feminine and neuter.
     
  21. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore

    It technically has gender, but for example, you note that "gender agreement" is not a huge stickler with learners. Also, all inanimate objects that would have gender in Old English (like say, masculine or feminine) are now treated as "neuter". The sheer majority of nouns in English would be neuter, no? ;)

    So I'm guessing that this last remnant of "genderness" might one day fall away.

    Mmm, the articles fill in the missing information, and if it is inadequate, then the words adopt a "strong" declension.

    In French, articles have become mandatory for denoting a quantity of anything (save something like beaucoup de, where it is implied) because the ending consonants are often elided so the articles are often the only way to tell number.

    Ah, but I am not a native speaker of Old Chinese.

    There are far more exotic reconstructions; IIRC, one of them is phregk.

    Also, the ancestors of any language often sound quite different. Would you immediately make the link between "Hrœdberð" and "Robert"? Aetaticum and age? Etc. ;)

    It's quite plausible to imagine when attempting to resolve the differences between Cantonese and Mandarin, since they sound different enough for to warrant a common ancestor that "looks" completely different.

    In some dialects, there are consonant clusters that Mandarin doesn't permit, as well, so their common ancestor would have had different phonotactics.

    Oops, I couldn't recall the exact construction, and pressed my brain too hard and invented something that didn't exist. :p

    Actually what I remember is something like toute ces explications; I was amazed that they could make plural agreement with "ces" but not with "toutes", before I realised that one had a vowel difference with its singular form and the other didn't.

    Ah, but wouldn't "all the" be declined? For example, French has "tous les", and in Latin, "totus" is declined based on the noun that it modifies. ;)

    Exactly. English doesn't really have case, except for a very limited amount of words (the pronouns). In French, you might enough information to distinguish indirect object from direct object (lui versus le/la) but word order is not very flexible either.
     
  22. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    For languages like Thai, I wonder if its gender system is perhaps somewhat like Vietnamese, where you might not have pronouns per se, but rather "social function words" just like Japanese honorifics. I think in Vietnamese you can use a third person construction (something like boy-pronoun loves girl-pronoun) to say "I love you"; the otherside has to invert the pronouns (e.g. girl-pronoun loves boy-pronoun). Is this a true gender system?

    Ah, it is indeed difficult to reconstruct; it take some considerable skill. I am afraid this thread will become even more off topic if I explain ;), but anyway ...

    Of major help are rime books -- these are basically poetry guidebooks that tell you which character rhymes with which character, among other thins, for ease of reference (and if you were stuck with a mental block). Sometimes, characters were used phonetically, to represent endings, initials, etc. Reconstructing pronunciations from rime books can be tricky; it is like a big logic puzzle.

    Then you also have several phonemic writing systems (note that all written languages are phonetic, but not necessarily phonemic) -- for example, when the Mongols ruled their empire, they expanded the Mongol alphabet to accomodate all the different phonemes of their new Empire. From them, we got a "snapshot" of the pronunciation Chinese at the time.

    You also have xiao'erjing, in which the Arabic alphabet is used as a sort of an Arabic pinyin for Chinese Muslims -- hence if your read historic texts in xiao'erjing, you will get the sounds directly rather than the characters. (Everything is inevitably translated into sound directly; too often people assume that Chinese writing is some esoteric ideographic writing system that can function without the use of sound.) Nuu Shu, a secret writing system used by women (in order to remain undetected by men) that was not revealed to the world quite recently, was also phonemic rather than syllabic.

    You also have comparative linguistics with the current dialects of China; just like we reconstruct PIE based on evaluating what is the most likely "original sound" through which regular sound changes can explain the evolution into its descendant languages, this can be used with Chinese. The lack of pronunciation information in writing can be a hindrance, and it changes the nature of reconstruction quite significantly, but it is quite possible!

    Then don't forget Japanese borrowings of Chinese characters, Korean hanja, and Vietnamese use of Chinese characters, which can reveal the phonology of the Chinese language at the time they were borrowed. I can't remember the name, but Japanese has an older system of borrowing Chinese characters that partially uses the Middle Chinese names for the characters. The very names "kanji" and "hanja" give away the different pronunciations of 漢字 at the different times they were borrowed.

    What's more controversial is reconstructing Old Chinese from Middle Chinese, since it is basically extropolating from already reconstructed data (hence, more room for inaccuracy). The flaw with reconstruction, as you might know, is that it is often based on probability. For example, suppose we had a fictional group of dialects and their words for "star":

    A: bo
    B: bor
    C: bo
    D; mo
    E: bo
    F; bo

    We might assume the parent word to be "bo", because it is easier to explain (probability-wise -- and note it is cumulative) the two sound changes that occur from "bo" to "mo" and "bor" than the four sound changes from "bor" or "mo" to "bo". But there is a slight chance (note it's not as though there is a 33% chance it wasn't "bo"; since it is cumulative it is somewhere closer to 1/36 or 1/216, etc.) that this was not what happened. And when you are reconstructing an entire lexicon, then perhaps making further reconstructions based on that already-reconstructed lexicon, especially if the reconstructions are based on rimebooks, this is where you might have some significant inaccuracies. Of course, this can be counter-stemmed by checking regular sound changes in general, like whether you have mu, mo, ma in one dialect, having the same meaning as bu, bo and ba in another dialect, but still.
     
  23. Whodunit

    Whodunit Senior Member

    กรุงเทพมหานคร
    Deutschland ~ Deutsch/Sächsisch
    I can't really tell you which gender is the most frequent in German! Maybe there have been some examinations upon this matter, but I don't know their solutions. The same might be true for Old English, but I don't know that. This would be a good question for our etymology forum. :)

    I must correct myself now. I misunderstood the words weak and strong in German. Thus, what is said about it in the Wikipedia about Old English is also true for German:

    Sorry about any confusion. Accordingly, I will have to edit my Old English example. ;)

    You would compare the words for "age" in all Germanic languages and see whether they could have a connection. English age and German Alter don't seem to have much connection indeed, and if you compare their roots, this would be right: The German word Alter is derived from IE *al-, which the English word age is an early (13th century) borrowing from the Romance languages (Latin aetas). If you now imagine that aetas and age mean the same, you could think that aetas went to aets, which got aage in Old French and finally developed to age in Modern English. The English equivalent to German Alter (adjective alt) is old.

    There's a current thread about consonant clusters, as you know. Would you mind mentioning some of them for a few dialects, if possible?

    And who says that toute ces explications is correct? It's still toutes. :p

    You would need to determine that. If you want adjectives ("all" is often declined like a normal adjective in those languages) to be declined, do it. It's obviously up to you if you invent a language. ;)

    Well, French word is a lot more flexible than the one of analytic languages, I guess:

    (1) Tu as téléphoné à qui?
    (2) A qui est-ce que tu as téléphoné?
    (3) A qui as-tu téléphoné?

    (1) J'espère que tu m'aideras.
    (2) Tu m'aideras, j'espère.

    They all mean the same, from which I think each sentence 1 is the most common. :)
     
  24. modus.irrealis Senior Member

    Toronto
    English, Canada
    Now I'm confused by what you mean by "gender system." :) But the Thai 1st person pronouns seem to work pretty much the same way the English 3rd person pronouns do in the sense that which pronoun you choose depends on the external non-grammatical context instead of grammatical properties of words.
     
  25. doman

    doman Junior Member

    Vietnam
    Vietnam, Tiếng Việt (Vietnamese)
    You're right !
    In Vietnamese, pronouns are always called as they are and without gender.
    For example, a mother say to her son "I love you", in Vietnamese will be "Mother love son", that son say "I love you too" will be "Son love mother too.", or a older brother tell to younger sister "Why didn't you hear me? "- in Vietnamese will be "Why didn't younger sister(place You) hear older brother(place Me) ? and ect.
    Pronouns in Vietnamese called exactly as its names and positions.
     
  26. kirsitn

    kirsitn Senior Member

    Oslo, Norway
    Norway, Norwegian
    In Norwegian we have both masculine, feminine and neuter nouns. The distinction between masculine and feminine is however optional in bokmål, one of the two official versions of Norwegian, whereas it is compulsory in the other version (nynorsk). We also have three third person pronouns - hun/han/det (she/he/it).
     
  27. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    Isn't the distinction between the two systems mainly that of orthography? Unless it's like French chatspeak where you drop gender and plural agreement if it doesn't affect pronunciation? ;)
     
  28. jfm Junior Member

    Sweden
    Let me throw in some hopefully useful bits into this discussion.

    1.
    INHERENT (GRAMMATICAL) GENDER is a property of words, e.g. the inherent gender of the German word "Mädchen" is neuter; hence "Das Mädchen".

    REFERENTIAL GENDER signifies something real or perceived in the object referred to. For example, when referencing "Das Mädchen" (grammatically neuter, semantically female) it is more common in German to use "sie" (grammatically feminine), not "es" (neuter), because in German, the feminine gender is strongly correlated with the semantic property "feminine".

    Hence the gender system in German is basically an inherent (grammatical) gender system, but it also employs referential gender in its pronominal system. In fact, inherent gender and referential gender are two, often contradictory, distinct grammatical processes, and they can be utilized simultaneously by any given language.

    Inherent gender may or may not be correlated with semantic parameters (or, defining features of the objects themselves). In many languages, the male/female opposition is very commonly the basis for grammatical genders. Animate/inanimate (or, human/animal/inanimate) is even more common universally. Less common are other types of semantic properties, e.g. "long thin things", "edible fruits", "domestic utensils", "abstracts", "paired body parts", etc.

    When semantic features other than male/female are used, the languages are commonly referred to as "noun class languages". However, the distinction between gender languages and noun class languages is merely one of terminology. Grammatically and functionally they are the same thing.

    Grammatical genders that are labelled "masculine" and "feminine" can be either weakly or strongly correlated with the notions "male" and "female", respectively. However, "masculine" is NOT the same thing as "male". The former refers to an inherent/grammatical property of words (linguistic units), while the latter refers to something outside language, namely, the "thing" itself. There no (known) gender languages where "masculine" (gram. gender) is semantically equivalent to "male" (biological sex). A masculine gender always includes non-male nouns and/or excludes some male nouns.

    While English has lost the use of grammatical gender, it still deploys referential gender in its pronominal system.

    The main point here is that there is a difference between inherent gender (a grammatical property of words) and referential gender (a semantic feature of the "object" itself). These may or may not correlate in any given language, and when they do correlate, they may do so weakly or strongly.

    2.
    It is not unusual for gender systems to simplify over time and even disappear. However, for some currently-not-well-understood reasons, grammatical gender systems are extremely stable. This means that, irerspective of degree of complexity, they can persist in languages over very long time periods without any major changes at all.

    They must have emerged in languages at some point, however. There are several current theories about how gender systems emerge, all of which involve restructurings/reanalyses of
    demonstratives, pronominals, and/or other word types.

    It is extremely important to keep in mind that simplification is not the only direction in which languages change. If it had been, we would have to assume that the first-ever human language was a super-complex, super-inflectional monster of a language. Otherwise we wouldn't be able to explain the complexities found in many modern languages.

    All serious theories about the origins of language assume a simple first language. At some point in its development, it grwe more complex by acquiring inflections, gender distinctions, case distinctions, tense distinctions, etc. Add to this issues of dispersal of speakers, parallel developments in different dialects, etc., and we get the global variation we see today.

    The main point here is (a) grammatical gender systems have emerged from somewhere, meaning that they were introduced into the language at some point in their development, and (b) if a gender-less language lacks evidence of prior grammatical gender, it probably never had any.

    ---
    jfm
     
  29. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    Hmm, actually it seems plausible to have the primordial language be highly inflected (in the sense of even more than Basque). If the human community size was relatively small and in one area (before they started scattering over the earth) then the linguistic community would be relatively stable, if not for tens of thousands to hundreds of thousands of years. It seems that languages that make frequent contact with other languages drop inflections (to accomodate each other), whereas a language whose environment is fairly stable would have the tendency to develop inflections. A highly inflected language is not necessarily a highly complex language, is it?

    I sort have this hypothesis that no language is really more complex than another -- if a language has an additional feature (take for example, Latin's inflection system), then it will be deficient in other areas (for example, Latin distinguishing only five vowel qualities). With the loss of many of the consonant clusters and the emergence of more restrictive phonotactics in Old and Middle Chinese, tones were borrowed. (One could also say the cause is the other way round - by borrowing tones [an additional feature] into the language, this provided extra information which allowed consonant clusters to be lost). People tend to say, "wow! This language is so hard!" but that's because the concepts are foreign, not necessarily that the concepts make the language more complex.

    For example, English has a simplified inflection system, but it became more complex in other things, like phonology, stress, word order and so forth.
     
  30. Nizo Senior Member

    Esperanto has no grammatical gender; except for words related to people or animals which may indicate their actual sex (viro/virino man/woman, frato/fratino brother/sister, aktoro/aktorino actor/actress), no distinction is made between nouns. The definite article is always "la" and there is no indefinite article. Adjectives, while they agree with nouns in noting the distinctions of plural and accusative, do not have any gender agreement. There are three third-person-singular pronouns: li (he), ŝi (she), and ĝi (it). There is only one third-person-plural pronoun: ili (they). Esperanto is very much like English when it comes to gender distinctions in language.
     
  31. Ilana New Member

    I'm not going to post long, but just make two small points:
    First, Norwegian have three sexes gramtically. Some parts of Norway (and some foreigners, for instance my hungarian grandfather) only use two, and use the masculine forms also for feminine. This sounds however a bit to "formal", and most people use all three sexes in everyday language.

    Second, I want to mention that also Hungarian only have one sex, like Turkish. Gramatically its not made any difference between male and female. I'm not trying to make an example now (since my Hungarian still isn't the best) but maybe some natives will find their way here..?
     
  32. jfm Junior Member

    Sweden
    In fact, it makes *no* sense at all. Where did it come from? How did it become complex?

    You're talking about a period we know virtually nothing about.

    No, that's not true. However, languages that are widely used as second languages (lingua francas, trade languages, languages of wider communication, etc.) have a *tendency* to simplify structurally. It not a necessary development, though.

    What is a "stable environment"? No language lives in total isolation. Every speech community lives in and encounters a variety of circumstances of a social, cultural, political, linguistic, whatever nature. Virtually anything may or may not have an effect on language structures.

    The development of inflections is a natural language process. It's part of something called grammaticalization.

    You may have in mind the difference between internal and external causes for language change. In brief, internal causes are such that are driven by mother tongue speakers, while external causes are driven largely by second-language speakers. These concepts cannot be equated with becoming more complex and/or simple. Even though external causes generally tend to have simplifying effects (especially on inflectional morphology), internal causes may result in either or both simplifications and more complex structures.

    Structural complexity is a relative value. It has to be compared to something. In terms of its morphology, an inflected language is certainly more complex than an isolating one.

    Latin may have a simple vowel system, but it's not deficient in any way.

    The underlying assumption is that the "amount" of complexity should always be the same, so that the total sum of complexity in any given language should always be, say, 100. The problem with this assumption is that there are no accepted methodologies for quantifying linguistic complexity, especially not if you intend to compare such different things as sound system and syntax.

    On what grounds can you claim that the amount of complexity lost in the English inflectional system has been balanced by an equal amount of complexity introduced into other parts of English?

    The thing about linguistic complexity is a curious thing. The claim that all language are equally complex is tacitly assumed by most linguists. However, no one has ever presented (a) any evidence for it, (b) any good arguments for why one should assume it to be so. What many people *really* mean when they say that "all languages are equally complex" is that "any language can express any thought". That is, not equal complexity, but equal expressiveness, which is a more important claim really.

    ---
    jfm
     
  33. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    I always thought inflections had a good chance of being pre-homo-sapiens. It would also convey an evolutionary advantage -- in hunting, for example -- to pack as much meaning into as little syllables/phonemes as possible.

    If the inflections were mostly regular and systematic (no doubt, we would also observe the tendency for the most used verbs to become the most irregular), it would have not been perceived as "complex" by its native speakers. It would have been natural. It would have been a system used for tens or hundreds of thousands of years, perhaps even partly biological.

    Perhaps the inflections were also signed, or were part-particle, part-inflections. For example, if you always appended a particle (let's call it "le") at the end to signify something that has passed, then perhaps it would become a regular inflection.

    I see no reason why "I attack him le" should be any less complex than "I attacked him". The latter is however more compact - it takes less time to say than the former. So we might see something like "I attack hleim", with such exotic ideas where the past tense information might be conveyed in the object, not the verb itself. (I'm not sure any language uses such a thing, just an imagination of how exotic constructions might occur.)
     
  34. Outsider Senior Member

    Portuguese (Portugal)
    Excuse me, but why have you always thought that?

    That's a biased comparison, though. A more balanced one would be between "I attacked him" and "I attack he le". And here's why the first is morphologically more complex: the lexemes "attack" and "he" have several versions (attack, attacks, attacked, attacking, etc.; he, him, his), whereas in the second (hypothetical) example they would not change (attack is attack is attack; he is he is he).
     
  35. karuna

    karuna Senior Member

    The planet Earth
    Latvian, Latvia
    Since no one has actually defined what it is meant by "complexity of a language" it is no wonder that there are no proofs for this. If we can equal "complexity" with "learning difficulty" then all empirical evidence shows that most languages, both inflexed and analytic, are equally difficult. English may seem simple to native speakers but not to its learners. For example, the rule when to use gerund or infinitive after the verb is very complex.
     
  36. Athaulf

    Athaulf Senior Member

    Toronto, Canada
    Croatian/Bosnia, Croatia
    I have no expertise in linguistics whatsoever, but based on my (minimal) knowledge of information theory, I can easily think of ways to quantify the complexity of language based on the information content per various units of speech, which can itself be quantified independently of what carries it. I have no idea what sort of research has actually been carried out along those lines, though.

    English syntax has a large number of extremely obscure and complicated rules, and these have undoubtedly become more complex and intricate as the inflections disappeared. Please also see the more general argument below.

    A very strong piece of evidence for (a) is the fact that children everywhere take the same amount of time and exposure to reach the normal level of proficiency in their native language. Furthermore, even adult learners have to take pretty much the same effort to reach fluency in different languages if we control for other factors such as closeness to one's native language, the level of motivation and intensity of exposure, etc. In analytic languages, it might be simpler to learn some very basic utterances, but above pidgin level, they're as hard as any others. It seems to me like these pieces of evidence are strong enough that it's the opposite claim that needs some proof to be considered seriously.

    I can also think of some quite plausible a priori arguments why languages could be expected to have roughly equal complexity, but I'm afraid we're already drifting far off topic...
     
  37. jfm Junior Member

    Sweden
    If you compare the case inflectional system of one language with the case inflectional system of another language, then yes, it is possible to quantify it, at least to a certain degree. However, if you intend to compare a case inflectional system with the phonological make-up of a language, then what would you propose as units of comparison?

    It is no problem devicing numerical quantifications of all different aspects of a language. Any aspect of a language can be converted to numerical values. But if you think about what those numbers are supposed to represent, the exercise might prove meaningless as you're not only comparing apples and oranges, but rather apples and feelings, i.e. ontologically totally different things.

    I'm not saying languages are *not* roughly equally complex. I'm only pointing out that it's a tacit unproven assumption, and cannot be treated as an axiom.

    It is often/sometimes useful to distinguish, or at least keep clearly in mind, the difference between form (structural properties) and content (semantics, function).

    On the content/functional side, there are no good reasons to assume anything else than that languages are indeed roughly equivalent. The point about children learning languages is the main argument here.

    In terms of structural properties, however, we have plenty of evidence that shows there are clear differences between languages. Some languages have more complex morphologies than others. Other languages have more complex phonologies than others. These things can easily be shown and even quantified.

    But to say that a language with a simple morphology makes up for that with a "comparable degree" of complexity somewhere else, like phonology or syntax, is an unnecessary assumption. It's often made, and there's nothing loony about it, but it doesn't actually achieve any analytic benefit.

    Take, for instance, a language that employs a complex noun class (gender) system, such as Swahili. We know that languages do not need noun class (gender) systems, so obviously a language can function without it, as does, for instance, Finnish.

    There are dialectal varieties of Swahili which show considerable simplification in their noun class morphology, spec. lesser number of classes/genders and lesser contexts where agreement is required. In short, the morphology is simpler than that found in Standard (Zanzibar) Swahili. Complexity hasn't been introduced anywhere else to make up for this. In fact, you'd be hard pressed to come up with any good reason why there should be. What would be the point or purpose of introducing complexity in a totally unrelated part of the language?

    Loss of inflections may or may not be accompanied by a more strict syntax. There's nothing strange with that, but there are no necessities involved. Loss of inflections do not by necessity mean that the syntax becomes more rigid. Many languages all over the world make use of grammatical categories not heard of in other languages. In Bantu languages we find, for instance, highly complex tense systems distinguishing between near past, remote past, immediate past, indefinite past, etc. These distinctions are not necessary for a language to work. They could easily be lost without diminishing the functional power of the languages concerned. In fact, while some Bantu languages have very complex tense systems, other Bantu languages have very simple ones.

    I have to point out that saying that a given language/dialect is structurally more complex or simple is NOT meant as a value statement. It is merely a statement about form, not content or the" functional power" of a language, i.e. its ability to convey messages, verbalize thoughts, etc. A structurally more complex morphology, phonology, or whatever, is not functionally superior in any sense.

    Neither are the simpler or the more complex languages more advanced in evolutionary terms. All living languages go through changes all the time. We have lots of textual evidence of Ancient Egyptian and Coptic covering a period of more than 4,000 years. That has provided linguists with much valuable data to develop theories about language change, and it is quite clear that in terms of at least morphology, languages go throw a pendulum-like process of simplicity-complexity-simplicity-etc, or inflecting-isolating-inflecting-isolating-inflecting, and so on. (There are other data to support this, not just Egyptian data.)

    ---
    jfm
     
  38. jfm Junior Member

    Sweden
    In structural terms, I attack him le is simpler than I attacked him.

    I attack him le = 1SG...VERB...3SG.OBJ...PAST

    In your example, assuming the words to be English, "him" is semantically complex. It consists of two semantic packets: 3SG (3rd person singular) plus OBJ (objective case). The other words seem to be semantically simple, i.e. one semantic packet equals one lexical realization.

    I attacked him = 1SG...VERB+PAST...3SG.OBJ

    In your second example, you have the same set of semantic packets, with the additional complexity of having structurally lumped the PAST tense information onto the verb.

    On a functional level, and taking your exampels at face value, they are equivalent. They convey the same information. Structurally, however, the second IS more complex. The ratio between semantic packets and lexical exponents is different.

    Additionally, although this is slightly more speculative, if we assume your examples to be from a language working like most languages work, then the inflected verb "attacked" would be an obligatory form, while usually tense information coded by free words or particles are not always obligatory, and they often (though not always) have a flexible position withint the sentence. In other words, there are less positional rules governing the distribution of a free particle than there normally is for an inflectional morpheme.

    ---
    jfm
     
  39. Athaulf

    Athaulf Senior Member

    Toronto, Canada
    Croatian/Bosnia, Croatia
    The mathematics of the information theory makes it possible to quantify information in a completely abstract way, totally divorced from what carries it. One could study the influence of various factors in phonology, morphology, and syntax on the entropy and information redundancy of various units of speech. If the final result is that the entropy of a logically self-contained unit of speech stays roughly the same as the language evolves, that's a pretty good indication that the loss of complexity in some parts has been offset by an increase in others. As far as I see, such a methodology would be immune to most of the objections that you raised in the above post.

    Of course, information entropy doesn't exactly correspond in any way to the intuitive human perception of "complexity", which is driven primarily by the impressions of the relative learning difficulty, but it certainly provides some insight into the situation. The whole issue should of course be addressed by someone with more knowledge about both information theory and linguistics than me (if it hasn't already).

    Are you really sure of this? Although I don't speak a word of Swahili, I would still gladly bet a monthly salary that some complications have been introduced to offset this simplification.

    I see at lease one very strong reason: information redundancy. Human language must maintain the same level of information redundancy to remain equally clear and easily understandable. If some complicated rules about morphological agreement simplify or disappear altogether and everything else remains equal, it will increase the entropy and decrease the redundancy of a typical sentence. On the other hand, if something else becomes guided by more complex rules, e.g. if syntax becomes more right, it will have the balancing effect of reducing entropy. Of course, there are also other ways of reducing the rate of information in order to enhance clarity (e.g. speaking more slowly), but these are considerably less efficient.

    Can you point an example of a language that lost inflections while maintaining the same freedom of word order? I can easily think of a bunch of counterexamples, from Bulgarian to Romance languages to English. I see no way how e.g. Croatian could lose inflections while maintaining the same freedom of word order. It would make most sentences totally ambiguous.

    And yet, every language has a few such "unnecessary" complications that give just about the same amount of headache to a typical foreign learner. They also tend to appear seemingly out of nowhere under specific circumstances (e.g Indo-European languages that lose noun cases tend to develop articles).
     
  40. jfm Junior Member

    Sweden
    If it did indeed stay the same, then yes, that would be a good and acceptable support for it. Without such a formula, however, there is no reason to assume either way.

    I can assure you, you would lose that bet. (See also below.)

    Language is not a mathematical formula. It is a social and communicative tool more than anything else. If a change occurs, it usually occurs because it satisfies a social, communicative or functional need. For instance, the lexicon of a language can, and does, easily enlarge without any necessary accommodations anywhere else in the language. In other words, the lexicon becomes more complex (by introducing new concepts, more refined semantic distinctions, whatever). This happens all over the world, especially when new technology is introduced, e.g. computers, cell phones, bicycles, cars, the wheel, etc. If the structure of the lexicon can become more complex, why not other parts of the language.

    Relating "information redundancy" to degree of structural complexity is ill-adviced. The level you should be concerned with is the functional capacity of a language, not its structural make-up. If some important and necessary semantic (functional) distinctions are lost, then it is those semantic/functional distinctions that are "made up for" elsewhere in the language. It is not by any necessity the degree of structural complexity that is preserved. If complicated case distinctions in the inflectional morphology are lost, then those same distinctions can be carried by the use of adverbials or prepositions instead. This does not require the introduction of a new word class, as both adverbials and prepositions may already exist and be used. It would be the functional capacity that is preserved (or shifted from one part of the language to another).


    Examples are plentiful. In sub-Saharan Africa, there are some 400-500 Bantu languages spoken. They all trace their historical origin to a proto-Bantu language spoken some 3000 years ago, so they are all "changed varieties" of a single (no-longer-spoken) language. If the equal-complexity-argument has any merit, then we should be able to see it clearly among these languages.

    Swahili has a relatively simple tense/aspect/mood-marking system. It distinguishes between a past, present, and future, plus a handful others, e.g. a "not yet" tense, a subjunctive, an imperative, etc. These are marked by inflections on the verb.

    Bemba is another Bantu language, spoken in Zambia. It's tense/aspect/mood-system is more complex than that in Swahili. It has not only one past tense but several, a "just now" past, an "earlier today" past, a "yesterday" past, an "earlier than yesterday" past, a "long time ago" past, etc. It also has several future tenses, plus subjunctives, imperatives, etc., etc. These are all marked by inflections on the verb just like in Swahili. In fact, the morphosyntax (the positioning of different types of inflections relative to the verb root) is identical in Bemba and Swahili.

    Looking only at the semantic make-up and its morphological realization, Bemba is much more complex than Swahili. It makes more fine-grained distinctions, and it consequently has more morphological markers (obviously).

    A sentence like "I worked yesterday" is possible in both languages, of course. While in Bemba it would sayable in one word (a verb inflected for person and the appropriate tense), in Swahili you would use the ordinary past tense form plus a sentence adverbial meaning "yesterday", almost like in English. In structural terms, the Bemba version is more complex. It requires more grammatical rules to produce and packs the same amount of information in lesser units.

    Now what about the adverbial. Isn't that an added complexity? In a sense, maybe. In Swahili, sentence adverbials usually appear at the beginning or the end of a sentence, though some flexibility does exist depending on how one wishes to emphasize, but the rules for their distribution are fairly straightforward.

    In Bemba, true adverbials are rare. In order to express what in Swahili would be adverbials (e.g. just now, sometimes, long ago), in Bemba you would either inflect the verb with an appropriate tense form, or you use a complicated relative clause. In fact, translating English sentences involving adverbials are way more complex in Bemba than in Swahili. (There are some adverbs in Bemba. However, their distribution is much more rule-bound than in Swahili.)

    Word order in general is otherwise more or less identical in both languages. In fact, on pretty much any level of analysis, Swahili is less complex than Bemba: morphology, syntax, and phonology. If Swahili and Bemba are equally complex, then Swahili should show more complexities on at least some levels, somewhere, but it doesn't.

    ---
    jfm
     
  41. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    Are there necessarily less positional rules? The occurrence of the position (and the ability for it to be moved) can itself convey another information complexity -- one makes up for the other.

    For example Singlish (which is really a spectrum ranging from pidgin among the elder adults to a creole among my generation), using a past tense form either signifies a more formal dialect (mesolectal, for example) or a stress on the past. Ruling out the former (in which past inflection is almost always used -- obligatory, as you say), a sentence like "went yesterday liao" and "go yesterday liao" can signify different moods.

    Some of these distinctions are unconscious (I myself did not realise that many Singlish particles had fixed contour tones assigned to each of them until it was pointed out to me) Sometimes it can be accompanied by a particle change ("went yesterday lor" --> nothing you can do about it).

    I am not arguing that each complexity or simplification directly leads to an opposite trend in another area. But for example, if you have obligatory inflection in one area, then that rules out any further complexity like distinguishing stress and mood based on the position or absence/presence of a particle. It becomes increasingly difficult to have multiple morphological complexities in one place.

    It seems natural for languages to drop distinctions it does not have to make, and to make distinctions -- irregular and inconsistent at first -- where the need serves them. I hear that many Basque speakers themselves have trouble with many of the complex inflections -- perhaps because they have begun using new constructions that render some of the inflections unnecessary (that over time becomes unused and discarded).

    Distinctions in AAVE between a sentence like "he be working" and "he been working" (a distinction that would not be made in standard English) seem to have arisen from a need to make distinctions in those areas, while simplifying morphology in other areas.

    I note that with "expanding a lexicon" may create more things to remember, but may not necessarily create many significant distinctions. Many Romance terms borrowed into English which we now use for specific senses ("sentiment", instead of mere "feeling") often replaced other words, rather than creating new distinctions or new minimal pairs. Whereas if you distinguish new vowels, or new morphologies (attacked instead of attack) new minimal pairs are created.

    Many dialectical argots will use obscure or archaic terms (in the standard language) as part of the regular lexicon, but not use other more usual terms in the standard dialect. IIRC, the theorised lexicon for the average native speaker of any language is generally around 60,000 words. Adding a scattering of new technological terms is insignificant. But what if you were to add 10,000 new words pertaining to various distinctions in technology? I bet many of these terms would be merged, just as many species of animals and plants that would have been recognised by different names in the past might be merged into wider labels.
     
  42. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    Phonological and morphological complexity are linked by the need to make distinctions in meaning (such as between "bat" and "pat", or "strike" or "struck").

    Perhaps it will turn out that our primordial language had a totally different nature, but I have imagined that it there was a low ratio of the amount of phonemes used to make morphemes. Hence, each vowel would have a meaning by itself, as well as each consonant, which might be uttered without vowels (or long strings of consonants uttered). So words might often be two consonants strung together.

    You would need to distinguish as many vowels and consonants as possible -- perhaps even vowel length, tone or so forth. This is so one can achieve the desired number of word distinctions in a language.

    But what if you were to increase the complexity in the methods to form words? Instead of just "lkhl-a", you might have "lkhla-lkhla" (redoubling) or lkhla-rghra, and eventually more complex combinations, like weiros or aetaticum. But now you would have plenty of ways to make new combinations -- all those vowel and consonant distinctions would become unnecessary, and eventually many of them will merge. It would have made sense for /m/ perhaps to have signified its own word in a primordial language because it is a very audible holdable consonant (and one of the easiest to produce). Yet now in many (most?) languages, the phoneme /m/ is no longer a valid word by itself.

    I note with Mandarin, as consonant endings got chopped off, creating an increasing number of homophones, Mandarin began to rely on bisyllabic words for concepts where other Chinese dialects would use one. For example, "renmin" (generally the people of a country, literally "person + people") is used instead of just "min" (people) to distinguish from the other homophones of "mín" (even with the same tone, other homophones have meanings that include a type of rock, a placename, "cord", and "multitude"). Earlier, with the loss of some consonant clusters from Old Chinese, tone became the replacing distinction in many Chinese dialects (and in the modern era, all) as well.

    Complexity in word formation can affect phonological complexity, and vice versa.

    As another side note to "obligatory" inflection, a problem in English that has been disturbing me is a sentence like "I didn't know it was Thursday already" (if for example, you have not been keeping track of the work week). Consistency would demand that the second clause (in the way it relays to "think") be the same tense as the first, but yet we know it *is* Thursday at the present moment. But "I didn't know it is Thursday" sounds slightly weird still. In this case, the obligation for inflection becomes a pain. In a language where time inflection would be optional, the second clause might be left uninflected in terms of pastness.
     
  43. Lugubert Senior Member

    Göteborg
    Swedish
    This view has some funny consequences. On the Swedish West coast, pojk(e) 'boy' would be masculine or common gender, ti(d)ning 'newspaper' feminine or common gender, depending on the speaker and if the word in question gets an article or an adjective. So, for pronouns, the animate gender is split in two. I can't right now think of any inanimate gender(neuter) nouns that are dialectally masc. or fem. Your method says that Swedish has four genders (pronouns han masc., hon fem., den common, det neuter), but standard Swedish distinguishes between two genders only (common, neuter) for nouns.
     
  44. Athaulf

    Athaulf Senior Member

    Toronto, Canada
    Croatian/Bosnia, Croatia
    I googled for some papers on the topic, but unfortunately I didn't find anything interesting. When I search for the relevant keywords, I get a bunch of engineering papers about language processing, but no papers with a theoretical analysis along the lines that I described above. Still, I'd be surprised if nobody has ever tackled that question; I'll try looking again when I find some time.

    However, does the lexicon of a typical speaker really increase? While there is no doubt that the size of dictionaries is increasing, and that certain people command a much larger vocabulary than others, I don't really see it as obvious that an average person's vocabulary has become larger than in the past. Sure, people know many words for modern technical gadgets and activities that didn't exist in the past, but they no longer know a vast number of words such as e.g. precise names of numerous plant and animal species, or technical terms used in agriculture, hunting, preparation of foods that are nowadays bought pre-made, making clothes, and whatever else we nowadays just buy instead of making ourselves. I'm not even sure that the vocabulary of educated individuals has increased, although the number of educated people is greater, and the sum of all particular professional vocabularies has swelled incredibly.

    However, my whole point was in preserving structural complexity regardless of any semantic functional concerns. What I have in mind is that human speech is an unreliable communication channel, and that structural complexity can serve as a sort of error-detecting and ambiguity-reducing code that increases the reliability of communication over this channel, in a manner analogous to error-detecting codes such as CRC in digital communications.

    Obviously, the structural complexity of any language could be greatly reduced without limiting its functionality at all. For example, take a language with a complex gender system and change all the nouns to be, say, masculine (in German, for instance, this could be done with almost no messing with the morphology and semantics, since there is nothing about a noun itself that tells about its gender). Obviously, anything can be said just as before, and the grammar has become much simpler, since all gender-dependent rules have been greatly simplified.

    However, what has changed is the average information redundancy per sentence. You can no longer identify a wrong or misheard sentence by the fact that gender agreement rules have been violated. In other words, the entropy has increased. Also, some sentences have now become ambiguous, whereas earlier it was possible to parse them in only one way that wouldn't violate the gender agreement rules. Thus, the probability of ambiguous and misunderstood sentences has increased. Assuming that all natural languages tend towards the same level of reliability in communication -- which I belive is a reasonable assumption -- such a change in a natural language would have to be offset by the emergence of additional rules elsewhere.

    But even if the necessary adverbials and prepositions already exist, this might result in the rules for their use becoming more complicated and subtle. Changes in syntax rules can be much harder to identify than the changes of morphology, which are always obvious. I can easily think of examples where complicated syntax rules have emerged with the same old words and word classes -- take for example the incredibly subtle rules for Spanish adjective placement, which was a non-issue in Latin.

    Thanks for the detailed examples, which I unfortunately cannot address with anything other than the above a priori arguments, being totally ignorant in the area of these languages. I guess that my perspective might indeed be skewed because I've never studied a non-European language, and as I've already said, these are just my personal musings, not educated opinions. Also, I can think of some examples along similar lines myself; for example, many Slavic languages have dropped certain tenses and aspects without anything else becoming obviously more complicated. But I'd really be curious to find out what actual research has said about the issues I've raised; for all that I know, the information-theoretical side of my argument might also be wrong, since I'm not an expert in that field either.
     
  45. john_riemann_soong

    john_riemann_soong Senior Member

    Singapore / United States
    English, Singlish, Chinese; Singapore
    Heh, this is an example that I attempted to use in my own post, but you did it more eloquently.

    Aye, this is what I tried to explain as well. While inflections are generally compulsory (though Singlish has special cases where inflection is used for emphasis), I notice there is a perceivable (though not necessarily obvious) change in what order a particle is used. (And there's also the case of how we feel compelled to use particles, even though our mood could be inferred without the particles being already there.)

    Tones are not totally totally necessary to meaning; if you knew the rough phonetics and you marred on the tones in Chinese, you could be comprehensible, though it would sound painful or funny. For example, there are no other meanings that you can interpret from Mandarin "renmin", so you could totally screw the tones on that one and still be understood. Even in cases where you could infer in other meanings, they often tend to be improbable (if you were given a choice between "you doing good, horse!" and "how well are things?", which would you infer?). It is only in certain cases that you would be truly ambiguous by screwing up tone.

    English stress rules and English rhythm are not critical to comprehension, yet they exist for some reason -- perhaps to help add redundancy and "error checking", as Athaulf suggests.

    Honestly, I had not thought of this before -- I had assumed another reason for observable equal complexity [though not necessarily prima facie quantifiable], mainly relating to the constant complexity of the "inherent brain language" of symbolic thought across the human race. I guess it could be both. If universal grammar appeals to you, I suppose -- I think we all really think the same mental language though we express it in different ways. (The complexity of the expression should therefore sum to the same.)
     
  46. MarX Senior Member

    Indonesian, Indonesia
    As far as I know, modern Persian, which is an Indo-European language, doesn't have any genders, either. It even has the same word for she and he.

    Personally, I find it unlikely that the she-he distinction in English will disappear. Although I notice that many non-native speakers muddle up she & he or sie & er when speaking English or German, even though their native languages have the gender distinction.

    As far as I know, Danish, Norwegian, Swedish, and Dutch distinguish the genders in the pronouns, but they generally don't have feminine-masculine distinction, although in spoken Norwegian and some Swedish dialects it is very much alive.

    I don't think the Austronesian languages ever had any genders historically.
    In fact, many other language families don't have any gender distinctions in the first place.

    Groetjes,


    MarX
     
  47. Lugubert Senior Member

    Göteborg
    Swedish
    We have, though, grammatical gender distinction in nouns, which for personal pronouns in the nominative case results in what may be regarded as four grammatical genders, for example Swedish han, hon, den, det (he, she, it, it), Dutch hij, zij, de, het.
     
  48. MarX Senior Member

    Indonesian, Indonesia
    Thank you, Lugubert, but I can't quite remember the word de used as a pronoun in Dutch. Does it mean it? In which context?

    Groetjes,


    MarX
     
  49. Lugubert Senior Member

    Göteborg
    Swedish
    Sorry, I was in a hurry and confused articles and pronouns.

    They have the unstressed hij, ze and neuter 't, stressed just hij and zij.
     

Share This Page