On Synthetic and Analytic Languages

The passage I wanted to talk about (see post below) was this one:

The injunction against split infinitives, for instance, is a consequence of the weird fact that English grammar is modeled on Latin even though Latin is a synthetic language and English is an analytic language.49

49A synthetic language uses grammatical inflections to dictate syntax, whereas an analytic languages [sic, at least in my edition] uses word order. Latin, German, and Russian are synthetic; English and Chinese are analytic.

For a start, I’d just like to make clear something David Foster Wallace doesn’t entirely (though he makes some amends in the next sentence). When he says “English grammar is modeled on Latin”, he isn’t meaning the grammar itself; he’s meaning so-called grammatical rules, like not having split infinitives or not ending sentences in prepositions. (Oddly, DFW is in favour of splitting infinitives (I’d noticed this already in the book and was preparing to pick him up on it), because he feels the rule’s derived nonsensically from Latin; but he’s against ending sentences with prepositions, presumably because he doesn’t realise it has precisely the same justification).

(Also – another aside – it is the fact that English “grammar” is modeled on Latin that causes the injunction against split infinitives, rather than it having anything to do with synthesis or analysis.) [ed. This is technically incorrect. It is the fact that DFW misuses the terms synthesis and analysis that led me to make this statement].

But anyway, I started wondering to myself about this distinction between synthetic and analytic languages, and when it comes down to it, I really don’t think there’s all that much in it.

For instance, the other night I asked a German friend, since she spoke a synthetic language, whether it was true she could construct sentences in any order in German, that word order could be played about with in this manner. – Being asked a question about an aspect of Germanness, naturally she became defensive: – her reply: “Well, you can play around with word order in English too”. (You see, Germans are just like us). I was prepared to concede this, since (as an idle “writer” who spends most of his so-called time “writing” merely re-ordering English sentences) I’d thought of it before, and tried to dismiss it as a point to come back to. – Anyway, we were drinking, and perhaps that wasn’t the best subject for a conversation anyway…

So I’d thought I’d try Latin instead. – I thought for once I’d avoid poetry (which is never the best for deriving grammatical rules) and have a look at some prose. So I took up some Cicero; – but found his rhetorical flourishes perhaps a little too extreme. I needed something simple and plodding. So I turned to Caesar’s Gallic War.

Here’s the opening of Caesar’s Gallic War

a) In Latin:

Gallia est omnis divisa in partes tres, quarum unam incolunt Belgae, aliam Aquitani, tertiam qui ipsorum lingua Celta, nostra Galli appellantur. Hi omnes lingua, institutis, legibus inter se differunt. Gallos ab Aquitanis Garumna flumen, a Belgis Matrona et Sequana dividit. Horum omnium fortissimi sunt Belgae, propterea quod a cultu atque humanitate provinciae longissime absunt, minimeque ad eos mercatores saepe commeant atque ea quae ad effeminandos animos pertinent important, proximique sunt Germanis, qui trans Rhenum incolunt, quibuscum continenter bellum gerunt.

b) In English

Gaul is a whole divided into three parts, one of which the Belgae inhabit, another the Aquitani, the third those who in their own language are called Celts, in ours Gauls. These all differ among themselves in language, institutions and laws. The Garonne river divides the Gauls from the Aquitani; the Marne and the Seine divides them from the Belgae. Of all these, the strongest are the Belgae, because they are furthest removed from the civilisation and humanity of the province, and merchants mix less with them and sell them less things which encourage effeminacy; and they are next to the Germans, who live across the Rhine, and with whom they continually waging war.

I’ve marked in bold where I’ve had to change around word order in English to make sense of the Latin word order (I’ve avoided noting changes of Latin phraseology, and those based on Latin’s superior succinctness – though this is derived in part, I guess, from its syntheticness).

Now, it doesn’t seem to me that there’s much of a difference there in the language’s supposed usage or not of word order as an indicator of meaning. The vast majority of the passage (and all Latin) can be rendered into English without changing the word order in the slightest.

It’s equally obvious that the words affected are always the same: – they are always the verbs, and the subjects of those verbs (and possibly the objects too – there just aren’t any examples here); – and this, of course, should be no surprise to anyone.

In fact, for someone claiming that Latin words can be in any order in a sentence, it is noticeable how many times in Latin prose (like in German, or so I recall) the verb is anchored at the end of the sentence (by which I really mean clause), with its subject somewhere close in the vicinity. – If word order doesn’t matter, then why?

The major exceptions to this are:

  1. The verb “to be”, which (as always) acts strangely: – in technical terms, the verb “to be” causes the subject and the object to be in apposition: – that is to say, the object is the subject; and in terms of a synthetic language, both are in the nominative case
  2. Verbs which are more for performance than actually part of the meaning of the sentence: – for instance, Cicero begins the speech I was looking at “Credo ego” / “I believe” – cf. this to the Spanish “yo creo que” / “I think that”, which is used much like English “Well … “, as a pause at the beginning of a sentence (listen to a Spanish footballer being asked – in Spanish, obviously – questions about the game he’s just played, and see if it’s not the first thing he comes out with).

What the difference between synthetic and analytic languages (or, at least, Latin and English) boils down to, I feel, is this: –

English tends to structure the cores of its sentences:

Subject – verb – object

Whereas Latin tends to structure them:

Object – subject – verb

[ed. This view needs more investigation!]

[The only real difference perhaps is the dropping of the accusative.]

All other parts of the sentences in both languages (in Latin, anything in other cases (though in Latin – AS IN ENGLISH – the genitive must always be next to the word it depends on / adjectives will be close by their nouns), in both all adverbs, prepositional phrases etc.) are (more or less) entirely moveable. Latin can – it’s true – create some nice effects which English can’t, particularly by a sort of adjectival wrapping – but there are (no doubt) unwritten rules limiting these too.

Perhaps, you will say: – but the difference is that Latin could structure its sentences in weird ways if it wanted to. – Yes, but the thing is, it tends not to, and I sense there’s a (grammatical) reason behind this. (My German friend’s answer to this was that, Yes, there were certain pretentious German writers etc.). Which brings me back to why I spend so much time as a “writer” fiddling about with the minutiae of sentence structure. It’s because, to me, for whatever reason, by whatever incomprehensible rule of the language, the sentence doesn’t sound right – and that this is something deep down in EVERY language, which is dependent upon word order.

(Oh yeah, there was a whole other section I was going to write – the opposite of this – about how English uses grammatical inflections absolutely everywhere – there just so familiar that you can’t see them. But I’ll save that for another time).


6 thoughts on “On Synthetic and Analytic Languages

  1. again, we seem to be encountering the same things independently. i was just reading about DFW a few weeks ago on this site, http://www.digitalemunction.com/2009/12/29/wtf-to-dfw-on-swe/. (sorry, i’m too lazy to code it).

    i always thought English was rather interchangeable with respect to clause order (analytic?), but that this didn’t give such tinkering any grammatical stamp of approval. it usually comes down to clarity. Clarity is what it usually comes down to. i mean, i can understand both ways…is this the sort of thing you think is meant by analytical? or is he just talking about translation, which your example rather proves for English as analytic, but says nothing about Latin’s sytheticity…new word!

    is clarity a rule of grammar (and not just style)? and if so, isn’t it relative to the reader/writer? where do we leave grammar and enter into style? doesn’t latin have a similar progression historically and demographically (for instance, all those Gauls certainly had their own Latin that we now associate with a particular period of usage)? is there any holdover from Germanic grammar in English? so many damn questions, do you want to purge this comment yet?

  2. Yes, there’s a lot more on this question – and many matters which are beyond my energy to investigate.

    No, he’s not talking about translation – he’s talking about the nature of the English language when he says it’s analytic. It is difficult to understand precisely what he means. (He also seems to make little distinction generally between word usage and grammar).

    There’s a possible example in that quiz, though:

    I’ve only spent six weeks in Napa.

    Which DFW thinks is wrong, and should be:

    I’ve spent only six weeks in Napa.

    This seems a reading of English where word order is imperative. – The trouble is, no one thinks twice about the first version of the sentence: anybody hearing it (apart from DFW) will naturally take “only” with “six weeks”, even though it’s not next to it. In fact, “only” just doesn’t seem to go with verbs as far as I can see. – If you put “only” anything else in the sentence, then it does indeed change the meaning:

    Only I’ve spent six weeks in Napa.
    I’ve spent six weeks only in Napa.
    I’ve spent six weeks in Napa only.

    Take another sentence:

    He only shot three Germans.

    What do you understand by this? Do you find yourself asking what else it was that he didn’t do to the Germans? – I’d be surprised. – So, in what way is what DFW is saying in any way to be considered “English usage” or SWE (Standard White English). Cos it ain’t standard.

    There is this whole thing in his essay where he pays lip-service to the idea that language is something constantly evolving, where words change meanings and usages – because it is something not possible to deny: – yet every argument he puts forward is to force language to remain static unnaturally (seemingly, because otherwise you wouldn’t be able to write dictionaries and books of proper American usage).

    And it seems all a bit rich to me, DFW marking students’ essays with lots of red ink over grammatical issues, when his own essays are packed with personal ticks which can hardly be considered SWE.

    Synthetic is a new term to me too. In fact, I looked it up on wikipedia. As you can see for yourself, synthetic seems to have a very different meaning to the way DFW wants to use it (or perhaps, DFW’s usage is just a subset of its general usage). The difference between the two (acc. to the wiki article on analytic language) is to do with using discrete words (in analytic) to indicate differences of meaning; while synthetic languages use inflexions.

    This latter was going to form part of a second “essay”. – In this essay, I am – I suppose – trying to demonstate that Latin is also fairly analytic (in DFW’s sense): it demands certain word orders (though this all needs more looking in to). In the next essay, I was going to show that English is also fairly synthetic (again, in DFW’s sense).

    I think clarity does come under grammar (as well as style). Contrary to the belief of many generations of schoolchildren, Latin is not constructed as a puzzle to be painfully elucidated; it was written for people to read in the simple manner we read English, and it was spoken – and for both of these to take place, there is a limit to what sentence reconstruction you can employ. Latin builds itself in sort of sense units, thus:

    Gallia est omnis | divisa in partes tres | quarum unam | incolunt Belgae | aliam Aquitani | tertiam (qui ipsorum lingua Celta) nostra Galli appellantur | Hi omnes | lingua, institutis, legibus | inter se differunt | Gallos ab Aquitanis | Garumna flumen | a Belgis Matrona et Sequana | dividit | Horum omnium | fortissimi sunt Belgae | propterea quod | a cultu atque humanitate provinciae | longissime absunt | minimeque ad eos | mercatores saepe commeant | atque ea (quae | ad effeminandos animos | pertinent) important | proximique sunt Germanis ( qui | trans Rhenum | incolunt) quibuscum continenter | bellum gerunt |

    Words are related to their neighbours; they are not structured without concern for word order. You can shift around these blocks to an extent within a sentence, and you can shift words to an extent within sense units – and there are certain wrapping effects Latin likes (where one sense unit becomes a subset of another) – but beyond this you can’t go far without someone looking at you oddly.

    I wish I’d been in DFW’s class, particularly around the age of 16. I’d have had a laugh, learnt little, and got some very bad marks.

  3. I must say, I thoroughly enjoyed watching DFW painting himself into a series of linguistic corners. I didn’t think the porn essay was boring, just very dry. However, I think you’re absolutely right, obooki: DFW pays lip-service to the idea of a living language being a work in progress, an organism that evolves. In fact, he’s as hide-bound a reactionary as any academic fusspot in matters linguistic.

    My two cents worth: clarity is everything. You write to be understood. If the current idiom is the more widely understood (as opposed to ‘correct’ usage as prescribed in Fowler etc), then go with the current idiom. There’s little point in attempting to write ‘for the ages’; there is no way of knowing what will sound ‘modern’ a century from now and what will sound hopelessly stiff and anachronistic.

    Write now to be understood now: the rest is nit-picking and tail-chasing.

  4. That footnote 49 is pretty infelicitously phrased.

    With respect to inflection (just one kind of “synthesis/analysis”), the synthesis of a “synthetic” language indicates that the ‘marshalling’ (syntax) of the words is done somewhat ‘by’ the words (via inflection). The analysis of an “analytic” language indicates that the words by themselves don’t point to their relation to each other qua their coherence in larger units (phrase, clause).

    You might think the figures of speech synthesis and analysis should be switched in the case of “languages” – but don’t think of them as pertaining to what the speaker/listener (or writer/reader) do. Think of “synthetic” as an activity internal to the words (as they appear in action) in a ‘synthetic language’, and “analytic” as denoting the discreteness, the atomic unchangeability, of words themselves in an ‘analytic language’.

    So: “Man bites dog.” “Dog bites man.”

    In a ‘synthetic language’, like Latin, you know what happened between that “man” and that “dog” regardless of the word order. In order to represent the action in the street (or to disclose the action, or whatever the relationship is between linguistic and material/phenomenal realms), the action in the sentence is ‘put together’ (syn – thesis) by the inflections of “man” and “dog”:

    Canem vir mordet. ‘Man bites dog.’

    Virum canis mordet. ‘Dog bites man.’

    The words in the first Latin sentence could be jumbled and still they’d mean ‘Man bites dog.”: vir mordet canem (Standard English word order.) Mordet canem vir. (Could be a question.) And so on: the word order is a matter of subtly (or not) communicating surprise, emphasis of some particular word, and so on. (In poetry, the word order also has to meet metrical demands – although, of course, those “demands” are also opportunities, valences of meaning.) But semantically, each inflected word (in such a simple sentence) does the work of its part of speech before (as it were) it connects with the other words in the sentence — the inflections “synthesize” the sentence. Hence: “synthetic language”.

    In English, it’s the word order, chosen by the speaker and grasped by the listener in accordance with established and competently anticipated conventions, that make intelligible ‘which bit whom’ or ‘who bit which’. “Bites man dog.”: just can’t tell ‘biter’ from ‘bitten’. The word “dog” is the same unit in both the exemplary sentences – you can’t tell from the word “dog” alone whether the dog in either sentence is ‘bitten’ or ‘biter’. So the uninflected words (that would be inflected if the sentence were in Latin) in an English sentence are already, and stay, the same (English) words they’d be in another, semantically different, sentence — as though the language had been (and always would be) ‘inwardly broken’ (ana – lysis) into its constituent words: an “analytic language”.

    (Of course, there is inflection of sorts in English – for example, the possessive genitive: “John‘s” – but compared to that in Latin, it’s tiny.)

    So, there’s a lot more flexibility, and more responsibility, for shepherding meaning by controlling the sequence of the words in a Latin sentence than in an English one that uses similar words to mean a similar thing: the English word order is more given, by living convention anticipated by competent users, to both speaker and listener.

    But — as you say, obooki, this relative increase of liberty/commitment in Latin word order is largely notional. Each Latin writer has a style, a pattern, that he or she favors and which prunes the notional plenitude of possible word orders of Canem vir mordet. to a small few (or one). And all the Latin writers, and, one guesses, speakers, follow basic patterns (like the one you’ve highlighted), just in order to be quickly understood.

    The other side of the coin: albeit with more words, you have equal ‘flexibility’ to say what you want in English – there’s no sentence in Latin that can’t be semantically reproduced in English at a skeletal level. (Of course, a translated sentence won’t mean exactly the same thing as its original, but that’s not a matter of “synthetic” vs. “analytic” languages.) The ‘extra’ words would be the way slight (or great) differences in emphasis, or whatever coloring of meaning you like, would be communicated in the English version.

    It’s an interesting thing – synthetic compared to analytic languages – but I don’t think it’s some tremendous impediment to bare-boned semantic understanding. ‘Poetry is what doesn’t get translated.’ -Frost — well, no, it doesn’t – not even in everyday language.

  5. (I’ll come back to the notion of English usage of inflection.)

    How about the Latin sentence:

    vehiculum iter coepit

    What does it means: “The vehicle began the journey”, or “The journey began the vehicle”? – I sense it must mean the latter.

    (And perhaps, in a similar way, in English we also feel, “The dog the man bites”, also means “The man bites the dog” – although we would like to introduce a relative pronoun).

    But neuter nouns in Latin raise an interesting question. If we’re merely concerned with nominatives and accusatives in terms of word order (and I’d argue we are); if we’re merely concerned with understanding the basic core of a sentence “X did Y”; then neuter nouns in Latin sort of bugger this up, because they don’t make any distinction between nominative and accusative.

    Why would a language introduce such lack of inflection for neuter nouns, if inflection were so important for understanding? It just seems perverse.

    Actually, where I’m drifting towards in all this is the question: why and how did we Europeans (seemingly with the exception of the Germans) change from using a so-called synthetic language to a so-called analytic language?

    I feel I’ve got some very simple answers.

  6. Yes, obooki, the analytic/synthetic distinction among languages is blurry at its clearest. Certainly in Latin, and (as you suggest) in English often enough, the horizon of convention (as opposed to ‘rules’ that operate independently of perspective) that includes a speaker and a listener in a conversation is a major enabler of any actual communication between them.

    But the fact of ambiguity – the experience of continuous disambiguation, negotiation, and conflict over meaning – doesn’t mean that this distinction is unreal or irrelevantly academic. Indeed, the systematic ambiguity in Latin of, for example, sentences with neuter subjects and neuter direct objects – an ambiguity that word order is supposed to address in English, where, much more often, there’s no way intrinsic to the words (as they appear) of distinguishing subject from direct object – makes clear specifically how inflected and uninflected languages differ — and (part of) what, for example, this English speaker is responding to in Latin poetry.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s