[Lingtyp] Distinction between semantic arguments and semantic adjuncts

[Juergen Bohnemeyer]

Hi Chao – I don’t know who introduced the term ‘oblique argument.’ It may well have been derived from or at least influenced by the model of ‘oblique case’. But oblique case does not play a definitional role in the notion of oblique arguments.

Also, there is obviously a great deal of variation in how different scholars use the term ‘oblique (argument)’. E.g., Dryer & Gensler (2013) use ‘oblique phrase’ expressly synonymously with ‘adjunct’. My own use corresponds closely to that in Role & Reference Grammar. For instance, here is Van Valin (1993):

“All arguments which appear in the core of a simple clause must be linked to argument positions in the LS [‘logical structure’, a representation of lexical meaning in terms of predicate calculus without variable binding; JB] of the predicate in the nucleus, and in the default situation, all arguments in the LS of the predicate must appear in the core of the clause. However, it is not always the case that an argument in the LS occurs in the core; in a passive construction, for example, the agent or experiencer, if overt, will be realized as an oblique constituent in the periphery. Among core arguments a further distinction is made between direct and oblique core arguments. This contrast is based on the morphological coding of the arguments: direct core arguments are those that are morphologically unmarked or coded with a direct case, as in dependent-marking languages like English and German, or are cross-referenced on the verb, as in headmarking languages like Lakhota and Tzotzil. Oblique core arguments are those marked by an adposition or by an oblique case. Thus English give has three core arguments; in Harry gave Bill the key there are three direct core arguments, while in Harry gave the key to Bill there are two direct core arguments and one oblique core argument.” (Van Valin 1993: 40-41)

In contrast, adjuncts are treated as modifiers, which in RRG means they are treated as expressing semantic predicates (Jolly 1993). This corresponds to conceptualizations in the Dependency Grammar tradition where modifiers are viewed as semantically and morphosyntactically relational, i.e., as opening argument positions for their heads, whereas arguments are dependents for which the head opens an argument position (Lehmann 1985). The Universal Dependencies database acknowledges this distinction, but does not follow it:

“The UD taxonomy is centered around the fairly clear distinction between core arguments (primarily subjects and objects) and other dependents. It does not make a distinction between adjuncts (general modifiers) and oblique arguments (arguments said to be selected by a head but not expressed as a core argument).” (link<https://universaldependencies.org/workgroups/core.html#:~:text=The%20UD%20taxonomy%20is%20centered,expressed%20as%20a%20core%20argument).>)

As to this point of yours:

“The question, however, is how to (semantically) rule out the tool used in a hitting action, for example, as a non-argument (maybe the tool in a cutting action has a different cognitive status than the one in a hitting action, but I’m not sure).”

This question is addressed head-on in Bohnemeyer (2007). In a nutshell, I argue that cut-type verbs require instruments that function like bladed instruments, whereas break-type verbs do not require any instrument and certainly don’t specify a particular type of instrument.

Best – Juergen

Bohnemeyer, J. (2007). Morpholexical Transparency and the argument structure of verbs of cutting and breaking. Cognitive Linguistics 18(2): 153-177.

Dryer, M. S., & O. D. Gensler. (2013). Order of object, oblique, and verb. In M. S. Dryer & M. Haspelmath (eds.), WALS Online (v2020.3) [Data set]. Zenodo.
https://doi.org/10.5281/zenodo.7385533 (Available online at http://wals.info/chapter/84, Accessed on 2024-07-09.)

Jolly, J. A. (1993). Preposition assignment in English. In R. D. Van Valin Jr. (ed), Advances in Role and Reference Grammar. Amsterdam: Benjamins. 275-310.

Lehmann, C. (1985). On grammatical relationality. Folia Linguistica 19: 67-109.

Van Valin, R. D. Jr. (1993). A synopsis of Role and Reference Grammar. In R. D. Van Valin Jr. (ed), Advances in Role and Reference Grammar. Amsterdam: Benjamins. 1-166.






Juergen Bohnemeyer (He/Him)
Professor, Department of Linguistics
University at Buffalo

Office: 642 Baldy Hall, UB North Campus
Mailing address: 609 Baldy Hall, Buffalo, NY 14260
Phone: (716) 645 0127
Fax: (716) 645 3825
Email: jb77 at buffalo.edu<mailto:jb77 at buffalo.edu>
Web: http://www.acsu.buffalo.edu/~jb77/

Office hours Tu/Th 3:30-4:30pm in 642 Baldy or via Zoom (Meeting ID 585 520 2411; Passcode Hoorheh)

There’s A Crack In Everything - That’s How The Light Gets In
(Leonard Cohen)
--


From: Chao Li <chao.li at aya.yale.edu>
Date: Monday, July 8, 2024 at 13:10
To: Juergen Bohnemeyer <jb77 at buffalo.edu>
Cc: <LINGTYP at listserv.linguistlist.org>
Subject: Re: [Lingtyp] Distinction between semantic arguments and semantic adjuncts

Hi Juergen,



Thank you for your detailed response. It’s clear that you use “oblique” in a restrictive sense to mean oblique arguments. In my mind, “oblique”, like the traditional use, refers to marked case marking (or marked flagging in an extended sense). I believe that’s why we find each other’s phrasing a bit confusing or problematic. By the way, it also shows how messy and murky linguistic terms are!



Back to semantic arguments, if go is analyzed as a three-argument verb, as suggested by your example, it creates some tension or at least some uneasiness when it is typically analyzed as a monovalent verb. Similarly, back to buy and sell, a four-argument analysis of them, whether standard or not, also creates some tension when they are often analyzed as or assumed to be bivalent verbs. This also brings us to prototypical two-argument verbs. In the literature on linguistic typology, hit, kill, and the like are generally presented as prime examples of such verbs. The question, however, is how to (semantically) rule out the tool used in a hitting action, for example, as a non-argument (maybe the tool in a cutting action has a different cognitive status than the one in a hitting action, but I’m not sure). An ideal solution to me is coming up with a definition of semantic argument that can successfully accomplish this and also match (more) with our intuition (e.g. intuitively the four participants involved in buy/sell are not equal in status).



(As a side clarificatory note not truly related to the content of your message, semantic arguments are not necessarily tied with verbs, as event-denoting nouns, for example, may also require one or more semantic arguments.)



Best regards,

Chao

On Mon, Jul 8, 2024 at 9:56 AM Juergen Bohnemeyer <jb77 at buffalo.edu<mailto:jb77 at buffalo.edu>> wrote:
Hi Chao – Yes, _buy_ and _sell_ have four semantic arguments (standard analysis since Fillmore (1982?) and Jackendoff (1990?) (citing from memory without having checked these specific works) and _cut_ has three.

But the phrase ‘formally realized as an oblique’ seems problematic to me. In my view, there is no clear morphosyntactic difference between obliques and adjuncts except that a proper subset of obliques are more or less obligatory and to that extent could be argued to be subcategorized for by the verb (or governed by the verb, in the traditional (pre-GB) sense of ‘government’).

But this obligatoriness should in my view not be taken to be a definitional property, because that would cut out the vast majority of obliques, which are not obligatory, yet are semantically clearly not modifiers. And that is, afaik, the whole point of distinguishing between obliques and adjuncts: to capture the fact that the latter, but not the former, are modifiers.

The primary difference between obliques and adjuncts is not morphosyntactic, but semantic. To be sure, it is never a happy outcome when semantic properties are needed to underpin morphosyntactic categories or vice versa. But the closer one gets to talking about form-meaning mapping, the harder it becomes to avoid such hybrid definitions.

Time location specifications are usually treated as adjuncts = modifiers. One exception (among a few) is _last_ as in (1). Locatives are usually treated as adjuncts = modifiers. One exception (among a few) is _live_ as in (2). Motion verbs commonly take path obliques, as illustrated in (3):


(1)          The meeting lasted from 9am until 10:20.

(2)          Sally lived in Buffalo in those days.

(3)          Floyd went from the Dean’s office to the library.

Finally, a puzzle: every event necessarily occupies a spacetime region. So (2) and (3) describe eventualities that we know must have occurred during particular moments in time, regardless of whether these are specified. So how then can we say that in those days in (2) is a modifier, whereas in Buffalo is an oblique? This is not at all a trivial problem. I could imagine treatments in which all time-positional and locative expressions are regarded as obliques. However, I would argue that is precisely the fact that all eventuality designators are by necessity compatible with spacetime specifications that limits the predictive power of the former for the occurrence of the latter.

HTH! – Juergen

Fillmore, C. J. (1982). Frame semantics. In The Linguistic Society of Korea (ed.), Linguistics in the Morning Calm. Soeul: Hanshin. 111-137.
Jackendoff, R. (1990). Semantic structures. Cambridge, MA: MIT Press.


Juergen Bohnemeyer (He/Him)
Professor, Department of Linguistics
University at Buffalo

Office: 642 Baldy Hall, UB North Campus
Mailing address: 609 Baldy Hall, Buffalo, NY 14260
Phone: (716) 645 0127
Fax: (716) 645 3825
Email: jb77 at buffalo.edu<mailto:jb77 at buffalo.edu>
Web: http://www.acsu.buffalo.edu/~jb77/

Office hours Tu/Th 3:30-4:30pm in 642 Baldy or via Zoom (Meeting ID 585 520 2411; Passcode Hoorheh)

There’s A Crack In Everything - That’s How The Light Gets In
(Leonard Cohen)
--


From: Chao Li <chao.li at aya.yale.edu<mailto:chao.li at aya.yale.edu>>
Date: Monday, July 8, 2024 at 07:39
To: Juergen Bohnemeyer <jb77 at buffalo.edu<mailto:jb77 at buffalo.edu>>
Cc: <LINGTYP at listserv.linguistlist.org<mailto:LINGTYP at listserv.linguistlist.org>>
Subject: Re: [Lingtyp] Distinction between semantic arguments and semantic adjuncts

Hi Juergen,



Thank you for all your information. If semantic arguments are participants predictable from the verb’s meaning (regardless of how they are overtly realized in real use), will buy and sell have four semantic arguments to you? Will cut have three?



(Also, were you defining obliques as semantic arguments (morpho)syntactically realized as an oblique? Or were you intending such semantic arguments as a subset of obliques as far as formal realization is concerned? If the former, what would you call those expressions that denote, for example, time or location but are also formally realized as an oblique?)



Thanks again,

Chao

On Sun, Jul 7, 2024 at 10:31 PM Juergen Bohnemeyer <jb77 at buffalo.edu<mailto:jb77 at buffalo.edu>> wrote:
Dear Chao – In theory, true adjuncts are modifiers, meaning they are not predictable based on the semantics of the verb. In contrast, obliques – semantic arguments that are expressed like adjuncts – must be predictable from the verb’s meaning. In practice, though, predictability is a matter of degree. For example, Koenig et al. (2008) show that English verbs form a continuum in terms of predicting an instrument. Regarding specifically verbs of cutting, Bohnemeyer (2007) compares the argument structure properties of such verbs across languages. – Best – Juergen

Bohnemeyer, J. (2007). Morpholexical Transparency and the argument structure of verbs of cutting and breaking. Cognitive Linguistics 18(2): 153-177.
Koenig, J.-P., G. Mauner, B. Bienvenue, & K. Conklin. (2008). What with? The anatomy of a role. Journal of Semantics 25(2): 175-220.

Juergen Bohnemeyer (He/Him)
Professor, Department of Linguistics
University at Buffalo

Office: 642 Baldy Hall, UB North Campus
Mailing address: 609 Baldy Hall, Buffalo, NY 14260
Phone: (716) 645 0127
Fax: (716) 645 3825
Email: jb77 at buffalo.edu<mailto:jb77 at buffalo.edu>
Web: http://www.acsu.buffalo.edu/~jb77/

Office hours Tu/Th 3:30-4:30pm in 642 Baldy or via Zoom (Meeting ID 585 520 2411; Passcode Hoorheh)

There’s A Crack In Everything - That’s How The Light Gets In
(Leonard Cohen)
--


From: Lingtyp <lingtyp-bounces at listserv.linguistlist.org<mailto:lingtyp-bounces at listserv.linguistlist.org>> on behalf of Chao Li via Lingtyp <lingtyp at listserv.linguistlist.org<mailto:lingtyp at listserv.linguistlist.org>>
Date: Sunday, July 7, 2024 at 21:12
To: <LINGTYP at listserv.linguistlist.org<mailto:LINGTYP at listserv.linguistlist.org>>
Subject: [Lingtyp] Distinction between semantic arguments and semantic adjuncts

Dear Colleagues,



The distinction between argument and adjunct is crucial for many linguistic analyses and much linguistic theorizing. However, how to define the argument and the adjunct and how to clearly distinguish between the two are controversial. Further complicating the issue is the existence of two related levels, semantic and (morpho)syntactic (e.g. what is semantically considered an argument may be (morpho)syntactically realized as an adjunct, as evidenced by passive formation in English). To be clear, this query is about the distinction between semantic arguments and semantic adjuncts (or semantic non-arguments). Specifically, in the case of verbs like buy and sell and in the context of a business transaction that generally involves a buyer, a seller, a transfer of goods, and a transfer of money, how many semantic arguments does each verb have, what are they, and what is the rationale behind the analysis? Similarly, in the case of cut, how many semantic arguments does it have and will that include the tool used in the cutting too? After all, isn’t the tool a necessary participant of the cutting action and how often do we cut something without using any tool? Likewise, how many semantic arguments in the case of bring?



Particularly, without looking at the different syntactic frames and constructions where these verbs occur or without paying any attention to how they are really used, on what (semantic) grounds can we say that Participant X is an argument of buy, sell, cut, or bring or that Participant Y is an adjunct of the same verb? For colleagues who’d like to make a further distinction between core arguments and peripheral arguments in addition to the distinction between arguments and adjuncts, then the following questions arise. Namely, if without looking at the different syntactic frames and constructions where these verbs occur, on what (semantic) grounds can we say that Participant X is a core argument, a peripheral argument, or an adjunct (of buy, sell, cut, or bring) and what is the difference between a peripheral argument and an adjunct?



Any input would be greatly appreciated. Thank you so much in advance for your time and help!



Best regards,

Chao
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20240709/6a4959da/attachment.htm>