Greek conversation
« previous post | next post »
Athanasia Chalari, "Greeks are ready to change", The Economist (Prospero) 10/24/2013:
Another interesting point about the difficulty in reaching a consensus has to do with social linguistics, how Greeks talk. […]
Greeks are very loud and they interrupt each other very often. The reason for that is the Greek grammar and syntax. When Greeks talk they begin their sentences with verbs and the form of the verb includes a lot of information so you already know what they are talking about after the first word and can interrupt more easily.
The reader who sent in this link noted:
Seems wrong to me–a quick look at WALS finds verb-first languages pretty even scattered over the world, plus many languages that pack more into the average word than Greek does its verbs, but I didn't have the time to test the claim thoroughly. But maybe you could do it, since you have a lot more information about turn-taking than I do. I was just skeptical that Greek is really that unusual in being 1) verb first and 2) relatively synthetic, so that one gets a lot of information out of the way in the first word of a sentence. (And all those verb-first, synthetic languages could just as easily lead to nice, harmonious exchanges of short sentences. I can imagine being more likely to interrupt if the crucial bits were at the end, since I would be inclined to say "get on with it!" or "You're wrong!" out of impatience.)
Dr. Chalari does not seem to have published any accessible papers on this subject. But there's a monograph that Penn's library doesn't have — Athanasia Chalari, Why Greeks Interrupt Each Other: The phenomenon of ‘overlaps’ in everyday Greek Conversations, 2012:
This book deals with the study of how Greek grammar and syntax make possible the occurrence of overlaps in specific locations, in everyday Greek conversation. Greek grammar and syntax have specific functions (such as freer word order, much information included in the verb, subject omission) which allow interaction to occur in specific forms (displays early projectability and early occurrence of overlaps). The co-existence of the above grammatical and syntactical characteristics within a Greek turn, make possible the appearance of overlaps in specific locations. Consequently, syntactic practices of Modern Greek language shape the organisation of overlaps that occur in everyday Greek conversations. In Greek, grammar and interaction organise each other and more specifically syntactic practices of language shape the organization of overlaps. This book proposes that if over time the suggestions of this study are generalisable then it could be supported that social and national stereotypes (for example that Greeks are loud or impolite or that they talk all together and interrupt each other) actually born in interaction.
There also seems to have been an unpublished 2005 version, "Why Greeks talk at the same time all together: examining the phenomenon of overlaps in everyday Greek conversations", with the following abstract:
In recent work on Conversation Analysis there has been a growing interdisciplinary interest among conversation analysts and linguists with regard interpretation of social interaction and grammar/syntax. This paper investigates the phenomenon of overlaps in everyday Greek conversation within the context of the amalgamation of interaction and grammar/syntax. The core of the present investigation would be the study of how Greek grammar and syntax make possible the occurrence of overlaps in specific locations, in everyday Greek conversation. Greek grammar and syntax have specific functions (such as freer word order, much information included in the verb, subject omission) which allow interaction to occur in specific forms (displays early projectability and early occurrence of overlaps). Furthermore, the co-existence of the above grammatical and syntactical characteristics within a Greek turn, make possible the appearance of overlaps in specific locations. Consequently, syntactic practices of Modern Greek language shape the organisation of overlaps that occur in everyday Greek conversations. In Greek, grammar and interaction organise each other and more specifically syntactic practices of language shape the organization of overlaps. This study proposes that if over time the suggestions of this study are generalisable then it could be supported that social and national stereotypes (for example that Greeks are loud or impolite or that they talk all together) actually born in interaction. Namely, those kinds of patterns (the production of early overlaps because of the Greek grammar and syntax) lead to those kinds of perceptions (that Greeks are loud and interruptive, as many of us characterise ourselves).
I haven't been able to locate any subsequent attempts to determine empirically if "the suggestions of this study are generalisable".
I once participated in some work on differences in overlap distributions across languages (or cultures), which notably failed to validate our pre-existing "social and national stereotypes" — Jiahong Yuan, Mark Liberman, and Chris Cieri, "Towards an integrated understanding of speech overlaps in conversation", ICPhS 2007:
We investigate factors that affect speech overlaps in conversation, using large corpora of conversational telephone speech. We analyzed two types of speech overlaps: 1. One side takes over the turn before the other side finishes (turn-taking type); 2. One side speaks in the middle of the other side’s turn (backchannel type). We found that Japanese conversations have more short turn-taking type of overlap segments than the other languages. In general, females make more speech overlaps of both types than males; and both males and females make more overlaps when talking to females than talking to males. People make fewer overlaps when talking with strangers than talking with familiars, and the frequency of speech overlaps is significantly affected by conversation topics. Finally, the two conversation sides are highly correlated on their frequencies of using turn-taking type of overlaps but not backchannel type.
We didn't have any Greek conversation to analyze, but based on our experience, it would seem to be appropriate to take a carefully empirical approach to the question of who does how much of what kind of interrupting, as well as what grammatical features might be correlated with possible linguistic or cultural differences in such things.
A sample of somewhat-relevant LL posts:
"The rhetoric of silence", 10/3/2004
"Crib notes and earphones", 10/4/2004
"Linguistic mens rea", 10/6/2005
"Hungarian speech rate and the tribunal of revolutionary empirical justice", 8/16/2006
"Next week: an experiment in primate communication?", 8/31/2008
"Conversational rhythms", 4/13/2009
"Two cultures", 3/20/2010
"The art of conversation", 7/24/2010
"Non-markovian yawp", 9/18/2011
"Marmoset conversation", 10/21/2013
"Speaker-change offsets", 10/22/2013
And I'll add one case where national stereotypes were validated, and another (noted by Ginger Yellow in the comments) where they were criticized:
"Nationality, gender, and pitch", 11/12/2007
"Thriving on confusion in The Guardian", 5/24/2006
Athanassios Protopapas said,
November 4, 2013 @ 5:45 am
Although I share the subjective impression that Greeks are loud and interrupt each other a lot, the whole idea of this being explained by some linguistic property sounds very strange (and empirically unsupported) to me. I don't know of any evidence that we start with a verb when speaking Greek in particular, at least not beyond what would be normal for any pro-drop language. Moreover, I don't seem to be any less rude when speaking English; somehow the phrase-initial pronoun fails to dissuade me from interrupting. It might have something to do with growing up in Greece.
GeorgeW said,
November 4, 2013 @ 6:28 am
Even if Dr. Chalari's claims regarding interruptions are correct (particularly that it is a function of the grammar, which I find dubious), the notion that Greeks are "impolite" is being expressed from the subjective point of view of a non-native speaker. I doubt if native Greek speakers would judge their conversational style as 'impolite.'
Elliott Hoey said,
November 4, 2013 @ 7:05 am
It's a seductive proposition to say that micro-interactional patterns scale up to institutional ones — and I'm sympathetic to the possibility that it's possible — but the examples Charlari gives in the unpublished paper "Overlaps in Greek" do not support what she is saying here about projection or societal discourse. Though she appears to understand what the issues are as regards grammar and interaction, the main issues with the paper (and presumably with the book) are her paying no attention to the particulars in which her overlaps arise. That is to say, overlap is specifiable formally, while an 'interruption' is culturally determined, and in order to say if something is an interruption you have to analyze it with reference to what sort of action the talk is implementing, the course of action it is situated in, and how it is treated by coparticipants.
VSO word order is very interesting psycholinguistically, and it's possible that it affords a recipient relatively more morphosyntactic resources for determining the likely trajectory of an utterance, given that more information will be available up front. In her paper though, it is not explicitly shown how initial verbs affect interactional practices in a way that is uniquely Greek. All of her examples can be explained in terms of action and action sequencing, which are universal issues for interactional participants; you have to very early on ascribe an action to someone speaking in order to begin planning a response. It's becoming clearer that turn-initial position is a general area for dealing with these sorts of issues. That is, you indicate the type of talk to come early on (Heritage 2013), meaning the effects Charlari points out for Greek are easily subsumed more general (and non-language specific) issues.
Graeme said,
November 4, 2013 @ 7:25 am
Fascinating topic but surely his conclusion is the weakest link of all. Why should pre-empting and livelier conversations necessarily militate against 'consensus'?
Ginger Yellow said,
November 4, 2013 @ 7:27 am
Only tangentially related, but it reminds me of Stewart Lee's claims about how differences in syntax help explain the differences between English/British and German humour, critiqued by MYL at the time here.
dw said,
November 4, 2013 @ 10:19 am
@Graeme: Dr Athanasia Chalari is a woman.
Jason Merchant said,
November 4, 2013 @ 10:30 am
First of all, modern Greek is not some fairly strict VSO language (like Irish for finite clauses at least): see Roussou and Tsimpli 2006 for some discussion. So any claims about relative order would have to start with a robust oral corpus analysis showing that VSO is relatively more frequent in the relevant cases than SVO, VOS, and OVS (all well-attested word orders in spoken and written Greek). Second, you'd have to compare languages that make a comparable number of distinctions in their inflectional morphology but are SOV (say, Greenlandic). Third, well, you'd have to control for a bunch of things. Let's just say I'm highly skeptical that such grammatical features are positively correlated in a significant way with turn-taking in a given language.
J. W. Brewer said,
November 4, 2013 @ 10:37 am
Everyone seems to agree that VSO is quite a common word order in Modern Greek, but lots of other orders are apparently not uncommon and not absolutely everyone seems to agree that VSO is the unmarked/default word order. But that actually suggests that given a large enough corpus of Modern Greek conversations, one could see if frequency-and/or-speed-of-interruption seemed to vary with the word-order (or conjectured attempted word-order . . .) of the sentences being uttered.
J. W. Brewer said,
November 4, 2013 @ 11:05 am
I had not seen Prof Merchant's comment when I hit submit, so I apologize for the partial redundancy. What would make a corpus inquiry like what I suggested both more difficult (because of the need to control for other variables) and potentially more interesting is that apparently some scholars of such things (plausibly, I would say, w/o really knowing enough to have an informed opinion) seem to think that speakers' choices in Modern Greek among the multiple syntactically permissible word orders for a given potential sentence are not random, but informed by various pragmatic considerations about information structure and the surrounding discourse context blah blah blah, all of which seem like factors that could themselves plausibly be correlated with the likelihood of being interrupted by an interlocutor. So if you had plausible theories that word order A should make early interruption more likely whereas pragmatic factor B should make it less likely but it turns out that ModGk speakers tend to use word order A in contexts dominated by pragmatic factor B, it might be quite interesting to see what corpus research actually showed.
Greg Morrow said,
November 4, 2013 @ 11:11 am
Pragmatically (if that is the right word), is Modern Greek new-information-first or new-information-last? Chalari's hypothesis is better explained if MG is new-information-first — once you have the new information, you are freer to overlap and interrupt. It's not about packaging the information in the verb, it's about what new is being communicated.
Jerry Friedman said,
November 4, 2013 @ 11:12 am
"In general, females make more speech overlaps of both types than males; and both males and females make more overlaps when talking to females than talking to males."
I can no longer restrain myself. In Isaac Asimov's story "My Son, the Physicist" (Wikiparticle), a male rocket scientist needs to communicate with astronauts who have an emergency around the orbit of Pluto or someplace. Light and radio waves take so long to get there that if he and the astronauts take turns, it will be most of a day between asking a question and hearing the answer. His mother saves the day by suggesting that they do what she and her woman friends do: talk at the same time. When I read this back in the '70s, I wondered whether it was sexist stereotyping.
Speaking of ethnic stereotypes, the scientist's name is Gerard Cremona.
[(myl) We should have been more careful of our generic plurals. Here's a graph showing differences:
How we interpreted the results:
Our analysis shows that it’s not that men are more likely to interrupt and overlap the others’ speech than women, it is that women’s speech are more likely to be interrupted and overlapped than men’s. This might shed some lights on explaining the contradictory results in the literature about who (men or women) tend to use more interruptions.
]
Vasha said,
November 4, 2013 @ 12:14 pm
Turkish is just the opposite: loads lots of information onto the verb which is usually last. And since Greece and Turkey have had so much cultural interaction… what?
[(myl) As Joyce Melton observes in the next comment, differences in interactional style are cultural rather than linguistic. English-speakers from different locations, classes, and/or ethnic backgrounds exhibit pretty much every kind of conversational style that exists, I think, or anyhow a very wide sample of them.]
Joyce Melton said,
November 4, 2013 @ 2:16 pm
I think it is much more likely to be simply a cultural phenomenon rather than linguistic.
It reminded me of being at a party in the 70s (in Texas) and watching a Missourian back a Californian all around the room. Having lived extensively in both places, I knew what was going on but neither person involved seemed aware of it. Missourians conduct conversations at much closer ranges, a foot or more closer, than do Californians, typically.
I don't think it had much to do with where they were going or what shoes they had on nor are Missourians more likely to wear glasses that do not correct for myopia.
alistair watson said,
November 4, 2013 @ 2:41 pm
Sorry if I've got the wrong end of the stick here, but it seems to me the interrupting by Greeks is culture-, not language-based. I live in Spain and one of the things that struck me here was how often Spanish people interrupt each other, or , if two people start talking at the same time, they both continue till they finished.As a Brit, I generally stopped on coinciding or being interrupted and found that my contributions to conversation became minimal. Debates here on TV tend towars shouting matches, as each side interrupts and counter-interrupts.
To prepare for my state teaching exams here, I had to study different elements of discourse, whcih included turn-taking as described by Schlegoff and Sacks. All of my Spanish students agree that turn-taking must be culture-based because they certainly pay no attention to its supposed tenets.
D.O. said,
November 4, 2013 @ 3:25 pm
Using telephone conversations for checking on interruptions has good and bad sides. Good, people have to use exclusively linguistic cues for possible interruptions (including grammatical structure, if someone is interested in that). Bad, timing for phone conversations maybe very much different because of the delays in the line. Also, it seems that frequent interruptions, especially in the beginning of the speech, are more expected in the face to face conversations were there is lots of non-verbal information.
Jerry Friedman said,
November 4, 2013 @ 4:27 pm
MYL: The graphs suggest that men do interrupt and overlap more than women, though that effect might be much less significant than the effect of the interruptee's sex. Is that what the numbers show?
Anyway, it seems that there is a real sex difference behind Asimov's story, though probably not enough to justify that the male engineers and astronauts don't see the solution but a woman finds it obvious.
Rubrick said,
November 4, 2013 @ 5:59 pm
People make fewer overlaps when talking with strangers than talking with familiars
Kudos to Mark for including corpora of witches talking to their cats; I gather access to those is rather closely guarded.
maidhc said,
November 4, 2013 @ 6:37 pm
Irish is a VSO language and English is not. If the hypothesis were correct, the Irish would interrupt each other more speaking Irish than if they switched to English. While I can't claim to have done a formal research study, I have observed the same people speaking Irish and English at different times, and I haven't observed any major differences in conversational style depending on what language they're using..
David Eddyshaw said,
November 4, 2013 @ 6:48 pm
Depressing to find this in the Economist, generally a beacon of fact-checking reliability in a naughty world, even in linguistic matters.
The actual thesis strikes me as slightly less solid than my own theory that people who put their verbs first (like us in Wales) get walloped by their neighbours (the heathen English) while people who put their verbs last try to conquer the world (Romans, Babylonians, Assyrians, Persians, Turks, Mongols … more of this and I might even start to believe it.)
Admittedly the Arabs are hard to fit into my scheme. I must demonstrate by science that Classical Arabic was underlyingly verb-final … must have been, the evidence is all there, what with all those Caliphates and all.
David Eddyshaw said,
November 4, 2013 @ 6:57 pm
(Obviously it cannot be coincidental that the Roman empire in the West declined in power just as the Latin language was shifting from SOV to SVO … it all fits, I tell you!)
David Eddyshaw said,
November 4, 2013 @ 7:16 pm
OK – I realise that I am on the brink of a grand synthesis here:
We verb-firsters, Welsh, Irish, Ancient Hebrews, Greeks (if you say so) are always interrupting each other. Obviously this cannot be conducive to military discipline.
Naturally those verb-lasters, Romans, Inca, Japanese, patient hearers-out of their superiors, will be able to organise vastly more successful military machines than we can ever hope to do.
QED.
Incidentally, the decline of the British Empire is clearly related to the decline of the teaching of (proper, Classical, SOV) Latin in our schools. How did I never see this before?
[(myl) Indeed a brilliant insight, with much more explanatory power than the previous attempts to relate martial prowess to the lack of nominal morphology.]
Adrian Morgan said,
November 4, 2013 @ 8:26 pm
I also studied cultural differences in conversation overlaps at university — as an undergraduate, not as a researcher — but in our class the interrupt-prone culture was French. No link to grammar was claimed.
Chris C. said,
November 4, 2013 @ 10:09 pm
@David — And let us not forget the Germans.
adriano said,
November 5, 2013 @ 3:01 pm
I wonder whether Italian-speaking Swiss are more prone to overlapping than their Southern neighbours (us). I'm no linguist, but I'm inclined to think that overlapping has more to do with culture than with grammar.
Patricia said,
November 7, 2013 @ 2:31 am
First, there really should be a "like" button on language log. Some of these posts are so humorous.
I think by "verb first" the author means Greek is a pro drop language and just has no terminology to express that other than to say that Greeks tend to begin their sentences with verbs.