Some notes on ditransitive verbs (part 1)
In his Outline of Classical Chinese Grammar, Pulleyblank (1995: 31) says that verbs of “giving, telling, teaching and the like” are ditransitive in classical Chinese, which he defines as the written language used roughly between 500 and 220 BCE. He states that in classical Chinese, the first object (OBJ1) of a ditransitive verb corresponds to the indirect object in English, with the second object (OBJ2) expressing the direct object.
This is indeed true for most sentences that contain a ditransitive verb, but not for all of them, or so it seems. I will first quote some of Pulleyblank’s examples, give some further examples, and then show there is at least one sentence which seems to do things the other way around. While writing this post it turned out ditransitive and indeed tritransitive verbs in the language of the oracle bones are complex enough to deserve a post in their own right, so that’ll be for next time. If only as a reminder that we should perhaps not complain too much about the vagaries of classical Chinese syntax…
Pulleyblank’s rule seems to hold…
First of all, some examples quoted and translated by Pulleyblank. Pulleyblank’s work does not include glosses, so I’ve included links to the relevant paragraphs in the Chinese Text Project (CTP) database. Its dictionary is still a work in progress, so for the odd place where its definitions do not seem to make sense, don’t hesitate to resort to a paper dictionary. Sadly, there are no good online classical Chinese dictionaries yet.
[…] 授孟子室。 (孟子，公孫丑下)
[…] to give Mencius a house.[…] 能與人規矩 […] (孟子, 盡心下)
[…] can give a man a compass or a square […]
For these examples certainly, Pulleyblank’s rule that OBJ1 expresses the indirect object and OBJ2 the direct object holds true. And indeed, it is not difficult to find more examples in the CTP database. A search for 與 yields, among others, the following:
子華使於齊，冉子為其母請粟。子曰：“與之釜。”請益。曰： “與之庾。”冉子與之粟五秉。(論語， 雍也)
Zi Hua being employed on a mission to Qi, the disciple Ran requested grain for his mother. The Master said, “Give her a fu.” Ran requested more. “Give her an yu,” said the Master. Ran gave her five bing. (tr. James Legge)由今之道，無變今之俗，雖與之天下，不能一朝居也。 (孟子， 告子下)
Although a prince, pursuing the path of the present day, and not changing its practices, were to have the throne given to him, he could not retain it for a single morning. (tr. James Legge)
…but there are exceptions
On the other hand, some passages seem to require a reading of OBJ1 as the direct object, with OBJ2 expressing the indirect object. Take a look at this sentence from the Zhànguócè 戰國策:
今王之地方五千里，帶甲百萬，而專屬之昭奚恤。 (戰國策, 楚策, 楚一, 荊宣王問群臣)
And now the King has a territory of over twenty-five million square miles, and a million soldiers, and he entrusted them all to Zhāo Xīxù. (Note that 地方 is not to be read as a compound. Wáng 2008: 111 explains the first phrase should be parsed as [[今][[王]之[地]][[方五千][里]]].)
Here, 之 is in OBJ1 position, but clearly expresses the direct object, and the OBJ2 昭奚恤 is undoubtedly the indirect object. So it appears that the rule Pulleyblank describes does not hold for all cases of ditransitive verbs expressing transfer of ownership.
It is possible that we are looking at a scribal error here. But perhaps there is more to it. 屬 does not seem to have been used as a ditransitive verb all that often. Perhaps this could be the reason for its unexpected behaviour, taking a direct object as its OBJ1 and an indirect object as its OBJ2.
Or perhaps it is used causatively: “He made it (OBJ1) belong to 昭奚恤 (OBJ2).” Does anyone else know of sentences where verbs that are usually not ditransitive are used causatively and take two objects? I imagine they would be few and far between, but would be interested to see any examples.
More on ditransitive and indeed tritransitive verbs in early Chinese languages next time. And a hat tip to Chris for pointing out the inconsistency in the Zhànguócè to me.
Pulleyblank (1995). Edwin G. Pulleyblank. Outline of Classical Chinese Grammar. Vancouver: University of British Columbia Press, 1995.
Wáng (2008). Wáng Lì 王力. Gǔdài Hànyǔ 古代漢語 [Ancient Chinese]. vol 1. Běijīng 北京: Zhōnghuá shūjú 中華書局, 2002.
Hi Daan,
great article. I was reading Walter Bisang’s seminal work, “Das Verb im Chinesischen, Hmong, Vietnamesischen, Thai und Khmer” (Vergleichende Grammatik im Rahmen der Verbserialisierung, der Grammatikalisierung und der Attraktorpositionen), and he mentions trivalent verbs in Classical Chinese two times, first, when talking about grammaticalisation paths for 給 on p. 63f and when talking about trivalent verbs on p. 92f.
One important reference here would be Alain Peyraube’s 1988 book, Syntaxe diachronique du chinois (évolution des construction datives du XIVe siècle auv J.-C. au XVIIIe siècle), as Bisang keeps referencing that book (and further articles on the issue by Chinese scholars).
The basic point of departure are the following three structures:
I. V IO DO
II. V DO 於 IO
III. 以 DO V IO
(the 給 type constructions start to appear in the Han era, with the structure V1 V2 IO DO, with a restricted set of V2, which later gives rise to the very productive pattern of co-verbs in Chinese, but this is irrelevant to the discussion at hand, and should probably be a topic for a blog entry of its own 😉 )
They do not mention any exceptions to this.
However, they also mention some tripartite verbs that are not verbs of giving:
Two citations from Wang Li’s textbook which we have been using for our online study group, are given by Peyraube:
抌兄之臂而奪之食
You twist your elder brother’s arm to rob him of his food
天生民而立之君
The Heavens gives birth to the people and sets them rulers (very hard to translate this literally)
This started a big debate about this in the 80s, with one camp just interpreting these as IO and DO, and another analysing 之 as an alternative to 其.
So we can conclude that usually verbs with two objects are verbs of giving, and follow a specific structure. (I might misunderstand you here, but what kind of “verbs that are usually not ditransitive are used causatively and take two objects” were you thinking about, unless you were talking about complements with 於 (it’s well known that local complements that usually stand with 於 can also omit it).
So I still think it must be a scribal error, though one think I would also check is to run a quick corpus study of the tendency of 之 to appear after any nonprominal objects (my hunch is that something like 屬昭奚恤之 would be dispreferred). So my question would be, does anyone know cases where 之 doesn’t directly follow the verb or preposition it’s governed by (cases of inversion excepted)?
Chris, thanks for sharing your thoughts. Looking at the examples in one of my books on oracle bone syntax here, it seems that in the language of the Shang verbs V DO IO structures do appear, although they are rare. I’ll give some example sentences with glosses in the next post, but as I don’t have a good dictionary here, I need to go to the library first.
About those verbs used causatively: I was indeed thinking of instances where 於 was omitted. That is to say, I think this sentence may originally have been intended to read 屬之於昭奚恤. What I would be interested to find out is whether the 於 has been omitted intentionally, or lost in the transmission process. Hence my question whether anyone knows of any other instances where causative verbs take both a direct object and an indirect object, rather than a direct object and a complement with 於.
Curious to see some stuff from the oracle bone syntax. However that would be a different language stage here and thus wouldn’t invalidate the point that by the time of the Zuozhuan, this kind of structure just does not occur…
I don’t see that there could be a causative vern with DO and IO, because for this you would need to have divalent verbs with IOs to begin with. Maybe locative verbs would count as they can usually omit the 於 of their directional complements, and if you causativised those, then you’d get that, in theory…