Event Structure (Part 2)
|Submitter Email:||click here to access email|
(Part 1 is posting 12.1389)
I am really overwhelmed by all the responses to the sum I posted on the 21st. I got so many replies with very interesting information. So I thought a second sum would be a good idea. Here it is, listing the replys that I received up to now in an alphabetical
orders of the contributors. The discussion is still going on but I'd already like to say a very big thank for the wonderful help I got!
I would like to contribute your research from Turkish. Fetch in means ''gidip getirmek which is to go and bring back.'' If you need further help from Turkish, am willing to supply it.Good luck!
Here is how Hmong speakers talk about fetching.
mus nqa ib khob dej rov los
go/take hold/one/cup/water/go~come back/come back.
fetch a cup of water
ib khob dej ('a cup of water') can be understood rather than expressed rov los can be expanded to rov qab los, where the collocation rov qab means, 'return' or 'go~come back'
I don't know much about this topic. But from what I see in the discussion, the English verb 'to fetch' does not seem to be composed of three event. According to Michael Swan, if it is good to say 'go and fetch' or 'go to fetch', it may mean that the 'go' event is only conventionally implied, not semantically included in the super event of 'fetch'. This is just what I thought. I may be wrong.
In Chinese, the 'go-take-come' has to be expressed in three verbs (serial verb construction) qu-na-lai.
One more note. In the question, the subevents are said to be: 1) going to some place, 2) taking something, and 3) coming (to the deictic center). Does it mean that the English verb 'to fetch' entails that after taking somethin g, the taker has to go back to the original place that is nearer to the speaker. For instance, can we say something like 'come and fetch it (yourself)', meaning someone from a place away from us to come to us to get something and than return (to the place he left). I ask this question, because in Chinese, the third event may be expressed by a different verb 'hui' (meaning 'return to
the original place').
Correct me if I am wrong.
AND IN FURTHER MAILS:
The differene between lai and hui in Chinese is that 'hui' presumes that there is a 'go/come' event and it means 'come/go back', while 'lai' only means 'come' (from a point further away from the speaker to the point of (or nearer to) the speaker. For instance, I live in Columbus, Ohio. I ca ll someone in Boston to go to London to get
something and bring it back to Boston. In this case, I use 'qu-na-hui' ('go-take-return'). If I want someone in Boston to go to London to take something but bring it to Columbus, I must use 'qu-na-lai' ('go-take-come'). I must point out that in Chinese, this is only a serial verb constructruction. The 'take' event c an be repaced by other events like 'buy' (i.e. 'qu-mai-lai': go to London t o buy something and bring it here) or 'borrow' (i.e. 'qu-jie-lai': go to London to borrow a book and bring it here).
Concerning your inquiry on the Linguist List. Russian distinguishes between two verbs: nesti (3Dbring), and prinesti (3Dfetch). The second verb is derivative with the suffix
'pri', though both verbs have the same root morpheme. The verb 'nesti' has the meaning 'to move holding something in hands' and it seems that apart from the three components you mentioned (1) going to some place, 2) taking something, and 3) coming (to the deictic center) there is one more: holding something, hence the similarity between German 'holen' and English 'hold'. Basing o n your classification I can say that the prefix in the second verb (prines ti) expresses meanings (2) and (3). Along with ''nesti' and 'prinesti' there is one more verb in the paradigm -'otnesti', which expresses meanings (2) and (1) (taking something and going to some place). Being a synthetic language Russian is inclined to express additional meanings by suffixes.
By way of a preliminary response, Siouan languages tend to handle this sort of thing with serial verbs. I am interested in this family especially in connection with Omaha-Ponca, but also comparatively. There's a general, but somewhat out of date discussion of Siouan
motion verbs by Allan Taylor in the International Journal of American Linguisti cs c. 1976 or so. It just summarizes the basic system, with some very incomplete mention of compounds of the strictly motional sort.
The basic system in Mississippi Valley and Mandan (together, Central Siouan) distinguishes eight stems representing the combinations of factors here/there, travelling/arriving, non-vertitive/vertitive. The other branches have fewer distinctions under arriving, or make them with a different set of stems.
The feature vertitive is referred to materials on other families as ventive or versive, and refers to travel homeward or to another starting point. In Omaha-Ponca the homeward sense is dominant. Some of the languages have developed ways to offer finer gradations in the here/there or travelling/arriving domains, e.g., here/to you/yonder or setting out/travelling/arriving.
Typical motion only compounds use combinations of arriving here/setting out or travelling there to indicate 'pass by here'. Other patterns of compounds combine motion + posture into forms like 'arrive and stand'.
In the Dhegiha and Chiwere languages there are an elaborate series of aorist (start/suddenly/repeatedly) auxiliaries that combine positional and motion roots (not quite the full set) with optional reduplication and causative. These are moribund in the more familiar Dakotan language.
Getting more into the area of your interest, senses like 'pursue' or 'bring' are generally compounded of 'having' or 'taking' + motion verbs. I do not specifically recall 'fetch' in Omaha-Ponca, but I know that some Siouan languages have 'fetchitive' formations involving transitivizing prefixes or proclitics (morphologized serial
AND IN A FURTHER MAILS:
So, I need to send you citations for:
Allan Taylor's IJAL article on Siouan motion verbs.
Boas & Deloria's 1941 Academy of Science grammar of Dakota.
Randy Graczyk's dissertation on Crow.
Allan Taylor's IJAL article on Siouan motion verbs.
Taylor, Allan R. 1976. On verbs f motion in Siouan languages. IJAL 42.4:287-296.
Boas & Deloria's 1941 Academy of Science grammar of Dakota.
Boas, Franz; Deloria, Ella. 1941. Dakota Grammar. Memoirs of the National Academy of Sciences 23 (Pt. 2), Washington, DC: GPO. There are several reprint editions. Try Amazon.com.See, e.g., pp. 84, 85, 86, 92-97, 101.
You'd probably also want to consult: 1971/1983. A dictionary of Teton Sioux, compiled by Rev. Eudgene Buechel, S.J., and edited by Rev. Paul Manhart, S.J., Institute of Indian Studies, U of SD, Vermillion, SD, and the Red Cloud Indian School, Inc., Holy Rosay Mission, Pine Ridge, SD.
Randy Graczyk's dissertation on Crow.
Try the list of Americanist dissertations at http://www.ssila.org.
Here is a Web reference to a summary on compound non-transitive motion verbs in Omaha-Ponca:
Note that dh in this is edh, and VN is nasalized V. The transcription is defined in the FAQ at http://spot.colorado.edu/~koontz.
Vern M. Lindblad:
My feeling is that fetch is almost always used paired with another word that can be analyzed as encoding one of the three subcomponents that you describe. The first pair that I am thinking of is go fetch where go3D1 and fetch3D2,3. The second pair that I am thinking of is fetch me where fetch3D1,2 and me3D3. I am not denying that fetch includes all three subcomponents, but I have a hunch that a corpus investigation would reveal that a very high percentage of occurrences include overt expression of either 1 or 3.
Sorry I haven't seen your post before, and I hope you don't mind my late commenting on the event structure of ''fetch'' in three languages I know. I have always had to grapple with this and other verbs in my translation studies courses with my students.
I have no references to give, but it seems to me that ''fetch'' is a complex event verb made up of three sub-events: (i) going, (ii) searching, (iii) bringing back. In English, ''fetch'' is packaged with all of them in spite of the existence of ''go and fetch,'' in which ''go'' expresses a sort of redundancy or overlap with sub-event (i). As a
confirmation, The New Shorter Oxford English Dictionary gives the following metalinguistic explanation for ''fetch'': ''go in search of and bring back.'' In French, the concept does not exist as such, i.e. not in the way it is packaged in English, and two sub-events are explicitly stated while (iii) is logically presupposed. As Prof. Khalifa rightly said, the French say ''va chercher'' (literally, go to search). In Arabic (MSA and Tunisian Arabic), however, it is (iii) that is packaged with (i) and (ii). We say ?aati (MSA) jiib (TA) (bring), which presupposes going and searching, and, of course, bringing back, but it is (iii) which serves as an umbrella concept for fetching. This is an instance of how languages conceptualise events by choosing which sub-events to foreground and background.
AND IN FURTHER MAILS:
Am I correct in assuming that in Arabic you literally say BRING, thus explicitly stating (iii) but intending more than that by presupposing (i) and (ii)?
[ZM] I am definitely positive about this. The verbs ''?aati'' and ''jiib'' (bring) are deictically further from the centre (i.e. the speaker), which assumes moving away from the deictic centre in search of sth, then back again to the deictic centre. So both verbs have a trajectory as a centre-periphery image-schema (i.e. where moving away from and to
the deictic centre are obligatory). However, verbs like ''?i?Tini'' (give me, in MSA) and ''?a?Tiini'' (give me, TA) presuppose proximity to the deictic centre, and no such image-schematic structure is possible.
A few additional comments:
In my kind of English (Midwest American), ''fetch'' is a word with restricted usage
1) used a command for dogs: Fetch!
2) dialectical usage: [Go] fetch me a cold beer from the fridge, will ya!
I would usually say:
3) Go get the book on the top shelf, will you please
4) Get me the book on the top shelf, please
5) Bring me the book on the top shelf, please
1), (2 without ''go''), (4) and (5) have three subevents with the emphasis on the bringing event.(2 with ''go'') and (3) have have two equal main events, with fetch/get having an implicit subevent.
Japanese has a different slant:
Sono hon wo totte kite kudasai
that book (obj) take (up) (and) come (polite command)
This covers the last two subevents and assumes the first. I assume that Korean would have a similar grammatic construction.
It's a little odd, to my mind, that no one has given Japanese, Korean, Chinese, Polynesian, etc., etc., examples.
It would be nice if a native speaker would respond to your query, but for what it's worth, the notion would likely be expressed in Japanese as ''tot-te k(uru),'' where ''tot-'' 3D ''tor(u)'' 3D ''take possession of,'' ''-te'' is a derivational suffix adding a meaning like ''and (temporal/logical) then'' and ''k(uru)'' is any appropriate form of the verb ''kuru'' 3D ''(will) come.'' There is also ''mot-te k-'' 3D ''bring'' with ''mot-'' 3D ''hold'' It works the sa me way: ''glom and come,'' but to my non-native intuition, there's less ''go'' implication.
I can't imagine ''go'' (''it-te'' with ''it-'' 3D ''ik[u]'' 3D ''go'') being included in the expression unless it was necessary to specify the location of the it em (reezooko-ni it-te biiru totte/motte kite kure ''Go to the refrigerator and fetch/bring me a beer'').
On the other hand, ''go to get/go fetch'' can be ''tori-ni ik(u),'' with assumption of intent to return but not actual return. That is, in the past tense, it would mean that the fetcher has gone and isn't back yet.
The equivalent with the verb ''mot(u)'' (pronounced approximately [motsu]) doesn't seem to work that way; ''moti [motSi]-ni ik-...'' makes as much/little sense as ''go to hold.''
Korean will work the same way; perhaps ''kaci-ko o-,'' but I'm out of practice. For some reason, I have apparently completely forgotten how to say such things in Mandarin and Hawaiian.
AND IN FURTHER MAILS:
(hon-o) tot-te k-uru
(book-acc) take-and come-will (3D ''[I?] will fetch [the book]''?)
Expressed events are of someone (the speaker or someone else--should be clear in context) taking hold of the book and coming to speaker's arena, both future.
Actual events include going (*probably* from speaker's arena) to where the book is, taking hold of it, and coming to speaker's current arena, expressed as intent or prediction of the future.
(hon-o) tot-te k-ita
(book-acc) take-and come-did (3D ''[I?] fetched [the book]''?)
Same events, expressed and actual, but already realized.
(hon-o) mot-te k-uru
(book-acc) hold-and come-will (3D ''[I?] will bring [the book]''?)
This time, the going-from-where-the-speaker-is-to-where-the-book-is is left out of the actual events; otherwise, it's the same as with ''tot-te k-uru.''
Ditto the ''k-ita'' version. I am not 100% sure that native speakers feel this as clearly as I do, though (I have just recently retired from 30 years of teaching Japanese, but I never attained native competence).
(hon-o) tor-i-ni ik-u
(book-acc) take--to go-will (3D ''[I?] will go fetch [the book]''?)
Expressed event, future: someone, perhaps the speaker, going towards the location of the book, with taking hold of the book not exactly stated, but given as the purpose of the going.
The understood/expected events include taking and maintaining hold of the book and returning to the speaker's locale. None of it has happened yet 2E
(hon-o) tor-i-ni it-ta
(book-acc) take--to go-did (3D ''[I?] will go fetch [the book]''?)
Expressed event: someone headed for the location of the book. Already happened.
Expected events are taking and maintaining hold of the book and returning to the speaker's locale, sooner or later.
By the way, my hypenation of the Japanese is not standard; I did it so I could do a morpheme-by-morpheme gloss.
I don't know where my Korean-English dictionary is right now. I checked my Hawaiian-English:
fetch--ki`i, lawe. Also ho`oki`iki`i, hali, hiki, hikihiki, ukiki`i; pukupuku mai (slang).
bring--lawe mai, hoo (properly ''ho'' with a macron on the ''o'') mai, pane`e.
''Lawe'' (and ''mai'') are the only words I remember learning (I took 2nd through 4th semester of Hawaiian in 1979-80). According to the H-E part of the same dictionary, it means ''To take, accept, carry, bring; portable; bearer'' (notice no ''fetch''). ''Mai'' means roughly ''hither,'' so ''lawe mai'' may be like the Japanese ''motte k-.''
Under ''ki`i,'' it gives ''To fetch, procure, send for, go after, attack; to seek for sexual ends.'' Sounds more like the Japanese ''tori ni ik-,'' but it will be hard to find a native speaker to verify any of my guesses. (Of course, sometimes when you list such things in a
summary, someone who knows better will come back with a ''that's all wrong--it really works like this.'')
One day I said I couldn't find my English-Korean dictionary. Then later, I found out the one I thought was E-K, but when I looked at it,it seemed to be K-E, was E-K after all! Well, if you can't follow that wild ride, it doesn't matter; what does is that I looked up
''fetch'' and ''bring'' in it, with interesting results.
''Fetch'' is given as ''ka-se kacye o-ta,'' where ''ka-'' is ''go,'' ''-se'' is one of several ''and''s, maybe one sort of ''and then,'' ''kacye'' is a form of ''kaci-ta'' ''to pick up, hold,'' and ''o-'' is ''come'' (let's pass the ''-ta'' off as the ''dictionary-form'' ending).
As you noticed immediately, this has all three events tied together! I frankly suspect that this is more a dictionary definition of ''fetch'' than what people really say, but I could very well be wrong about that.
''Bring'' was defined the same way, without the initial ''ka-se'' 3D ''go and.'' There were some one-word glosses the words as well, but I can't evaluate their event structure. By the way, I'm using Yale romanization, one of several ways to write Korean with latin letters.
Another taste of data to chew...
Dennis McConville :
I thought it might interest you if you had some information from a different language - in this case Thai - although my knowledge of that language is not very good. However, good enough to be able to discuss the matter with my Thai wife.
Holen or fetch is rendered in Thai as (my transcription): /pai lenamah/
literally 'go and bring'. Interestingly though there is also an other expression: /pai ow/ literally 'go get' and /pai rap/ literally 'go collect' or 'go pick up'. Interesting since only one requires 'and' between the events.
I rather suspect that something similar happens in other S.E. Asian languages but I am not familiar enough with them.
ich habe gerade die Summary B8ber 'fetch' in der Liguistlist gesehen. Ich wB8rde sehr gerne wissen, was bei der Untersuchung rauskommt. Ich arbeite an 89hnlichen Problemen, aber mir geht es eher um die Frage, was fB8r Ereigniskombinationen eine VP (also nicht bloFE ein Verbstamm) ausdrB8cken kann, und wie sich das in der Argumentstruktur niederschl89gt. In hoffe ntlich wenigen Tagen werde ich ein abstract zu dem Thema fertig haben. Ich erlaube mir, ein paar Gedanken aufzuschreiben zum Thema 'fetch' usw.
(Vielleicht ist manches schon bekannt.)
Neben den von versch. Leuten erw89hnten Konstruktionen vom Typ 'go and fetch it' bwz. 'go to fetch it' hat English so eine Art serial verb constructioget it, go find it, go look, go fetch it
M.W. gibt es das nur mit 'go' als linkem Verbstamm. Die Konstruktionen sind sehr interessant, weil sie in keinen Kontexten vorkommen, wo (overte) Flexionsmorphologie erforderlich ist. Es gibt keine solchen Flexionsprobleme im Franz88sichen bei 'aller chercher' und im Deutschen:
wir sind eine rauchen gegangen.
sie geht arbeiten
Vielleicht sind die englischen Konstruktionen linksk88pfige oder exozentrische Komposita, aber das ist nocht keine Erkl89rung fB8r die Flexionsprobleme, weil in parallelen F89llen im nominalen Bereich die Flexion zul89ssig ist: 'bailouts, sister(s)-in-law(s)'.
Ich hatte mir B8berlegt, ob man das erste Teilereignis bei 'fetch' (also die Fortbewegung vom Ausgangsort) unbedingt in der lexikalischen Bedeutung vom Verb festschreiben muFE; vielleicht k88nnte man es als Implikatur ansehen. Aber das kann bei 'fetch' nicht stimmen
(u.a. weil sich 'bring' und 'fetch' im (Nicht-)Vorhandensein der vermeintlichen Implikatur eindeutig unterscheiden, und sonst an kaum etwas); 'holen' ist schwieriger. Man kannz.B. 'etwas aus seiner Tasche rausholen/hervorholen' usw,, wo kein Fortbewegung des Agens notwendig ist, aber bei 'fetch' geht das definitiv nicht. 'fetch' kann aus
irgendwelchen GrB8nden keine direktionalen PPen an sich nehmen (*'she fetched it to/from the office') (wobei 'fetch it back' akzeptabel ist ('back' ist glaub ich daran schuld, weil es sich in anderen Possessivkonstruktionen ein fB8r Richtungsangaben untypisches
Verhalten aufweist). 'holen' l89FEt die Neutralisierung des 1. Teilereignisses genau dann zu, wenn eine direktionale PP als Komplement erscheint ('er holte es' ist m.W. genau dasselbe wie 'he fetched it' (auFEer daFE 'fetch' stilistisch markiert ist: die
H89ufigkeit ist gering, und 'fetch' wirkt altertB8melnd-gehoben-britisch, auFEer vielleicht wenn man einen Hund anspricht. Als Muttersprachler des (australischen) English verwende ich 'fetch' so gut wie nie, ich wB8rde viel eher sagen '(go (and)) get it'.
Man muFE dieses Stilproblem bei der Informantenbefragung in betracht ziehen, und sichergehen, daFE die Leute die S89tze nicht wegen des Stils ablehnen.)
Fiona Mc Laughlin:
The notion of ''fetch'' in Wolof (Atlantic, Niger-Congo) is expressed by the verb 'jCEl' to take/get in combination with a distal verbal suffix '-i' indicating a direction away from the speaker.
im Koreanischen wird die Situation ''holen''/''fetch'' so wie ''gehen-nehmen-kommen'' konzipiert.In anderen Worten bildet man einen komplexen Satz, um eine solche Situation auszudrB8cken. Ein solches Thema wurde in dem Aufsatz von Lehmann & Shin & Verhoeven
2000 ''Unfolding of situationperspectives as a typological characteristic of languages'' STUF 53, 1:71-79 behandelt.
Another language that breaks it up: Japanese uses totte mairu, (
Herbert F. W. Stahlke:
Finding a Yoruba equivalent for ''fetch'' depends on choosing an argument structure. For
He fetched me home a book from school.
Yoruba will have a verb for each argument. (Since I can't show tone on vowels in this editor, I'll put tone marks after vowels, ' for high and ` for low, unmarked for mid.)
o' lO si' ile'iwe' mu' iwe' wa' si' ile' wa'
he go to school pick-up book come to house come give me
There is argument as to whether si' ''to'' an fun' ''give'' are verbs or prepositions, argument that is so far inconclusive. How this breaks down for event structure is a separate question, but it contains at least four verbs, six if you count si' and fun'. There's been a lot of work on verb serialization in Yoruba, and you might look at some of
Mark Baker's for a formal syntactic approach. In the 70s there was lot of less theoretical descriptive work by people like Bamgbose that is worth looking at as well.
Hi. Just saw your post on LINGUIST. I've been working on Yahgan, an isolate from Tierra del Fuego and the southernmost language of the world. The dictionary I'm re-editing contains 32000 words, the vast majority of which n either as such or also interpretable as nominalizations in the right contexts.
Most of these are multi-root, serial forms. Roots can occur optionally with valence-shifting prefixes. The language is SVO, as are most serializing languages. Incipiently, the language (when still robust during the 19th Century when the dictionary was compiled) was headed towards what is referred to as ''bipartite'' stem structure as found in many North American languages- including Atsugewi as discussed by Talmy. This includes an instrumental/bodypart prefix complex, stative/adjectival central root complex, and finally a locative/pathway complex.
Flanking this central eventive core are on the left, the primary valence shifting elements (less optional here), and on the right, aspect/mode/tense suffixes.
The language was actively grammaticalizing presumably from an isolating/analytical antecedent form (as one finds also in Tibeto-Burman languages, such as Lahu (see Matisoff))- the valence shifters appear to be coming directly out of the instrument/bodypart set (similarly one finds that in Hokan Pomoan, for instance, that
bodypart/instrument terms have causative and other values).
The aspect markers are transparently derived from pathway/locative elements, including forms for ''sit/stand/lie/jump''-as found in many other serializing languages- the metaphor for continuity of process is degree of surface contact- the deepest such contact is ''lie'' (really ''lie within'', as in sleeping snugly with your bedclothes- English has similar ideas in ''being wrapped up in'' some process, ''involved'' etc (lit/fig Latinwise). Sitting involves split attention but more towards the process in question, standing more vigilant to some other matter potentially arising, and jumping off equals complete disengagement from the substratum, and thus sudden end to the continuity of the process (English ''coming to the end of'', etc.).
Roots (and thus also the incipiently prefix and suffix elements) come in several flavors- there is massive split for inherent singularity or plurality (in other words, etymologically different forms describing the ''same'' subevent). As grammaticalization was progressing, something interesting was happening- some of the singular or plural inherent
forms dropped out, and new affixed forms were taking on generic plurality or singularity functions- still more or less classificatory depending on the original semantics of the root, so not generic plural/singular morphology, at least not yet. In addition, further
developments included duals (mostly with animates), specific trials, quartals, etc. Each of these forms is transparently related to the numerals.
In addition, the root set was also divided, in many cases, into forms describing the ''same'' subevent when descriptive of participants of one's own, or belonging to someone else. I've not found this anywhere in the literature, but suspect its there in other languages- perhaps related, for instance, to deictic point-of-view. ''Same subevent'' forms also split on whether action is done broadly (in time/space) or narrowly (also interpretable as short/fat or long/thin, etc.). The system is so predictable, in fact, that it creates a sort of phonosemantic system, though one completely different from that found in ideophones or expressives, though ultimately relatable to them.
I don't yet have enough connected discourse available to work out such things as switch-reference, or other interesting stuff related to serialization, but nuancing relating to the stringing of valence shifters on succeeding roots may be part of it.
The overall time-sequence structure of verb stems seems clear-cut. Only the interpretation is left to tweak- this probably depends on which root is highlighted- and since accent and other prosodic structure is unmarked its hard to determine from the data at hand, except from the phonetics than nothing).
Since the language had to have come through an isolating/analytical stage at some point in its typological past to get where it is now, it would be interesting to look the development of such languages from the other end, to see what kinds of attrition or addition was going on lexically in order to create such a coherent system dividing up the
eventive universe into such quantized chunks for later recombination. We know that as productive morphological derivation breaks down increased lexicalization is the norm (look at English, for instance, or many other IE languages as compared to the parent
PIE). Could it be, then, that the ''stew'' of forms is far from arbitrary as it works out its kinks to become a new, coherent system? Work on concept hierarchies (Eleanor Rosch, for example) might come in here, but I don't get the impression folks here are looking at overall integration, at least not yet).
Thanks for the interesting summary of your inquiry.
In Chinese (the numerals are tone markers):
/qu4/ means 'go, or move away',
/na2/ means 'take/hold by hand',
/qu3/ means 'take & come',
/lai2/ means 'come'.
So ''fetch obj'' is either /qu4 na2 obj lai2/ or /qu4 qu3 obj/. If one says /qu4 na2 obj/, it means to go and hold the obj, without indicating the resultant direction (come back or go away with the obj).
|Original Query:||Read original query|
Sums main page