FREE ELECTRONIC LIBRARY - Thesis, dissertations, books

Pages:   || 2 | 3 |

«I. Introduction Consider sentences like (1): 1. Null Complement Containing Sentences a. Aryn followed b. Marie-Odile promised c. Corinne left d. ...»

-- [ Page 1 ] --

The Semantics and Syntax of Null Complements*

Marie-Odile Junker1, Robert Stainton2 and Catherine Wearing3

Carleton University, 2University of Western Ontario, 3Wellesley College

I. Introduction

Consider sentences like (1):

1. Null Complement Containing Sentences

a. Aryn followed

b. Marie-Odile promised

c. Corinne left

d. Samir found out at midnight

e. I applied

f. They already know

g. He volunteered

h. Abdiwahid insisted

i. I suppose

j. Paul gave to Amnesty International These illustrate the phenomenon of null complements -- also called ‘pragmatically controlled zero anaphora’, ‘understood arguments’, and ‘linguistically unrealized arguments’. In each case, a complement is (phonologically) omitted, yet (a) the sentence is well-formed and (b) the meaning effect is as if a complement were present. This contrasts on the one hand with structures that lack complements, but are ill-formed as a result – e.g., (2a-c) – and, on the other hand, with structures that lack overt complements, are well-formed, but do not exhibit the meaning effect of a complement – e.g., sentences (3a-b).

2. Contrast with Ill-formed Structures a. *Aryn purchased b. *Marie-Odile guaranteed c. *They already expect * We are grateful to Corinne Iten and Aryn Pike for discussions on an earlier draft. Funding for this project was provided by the Canada Research Chairs program, Carleton University, and the Social Sciences and Humanities Research Council of Canada.

Carleton University Cognitive Science Technical Report 2006-03. http://www.carleton.ca/ics/TechReports

3. Contrast with No Complement At All a. He died b. Paul laughed Throughout this paper, our focus will be on the nature of null complements, including especially the semantics and syntax of null complement containing sentences.

There is another issue about null complements which interests us, namely what licenses them: what features of a verb determine whether it will allow null complements or not?

But that question is addressed in another paper (Iten et al., 2005).

We begin with a very general characterization of the phenomenon. We assume in what follows that verbs have semantic frames, which specify the linguistically inherent participants for that verb. By a linguistically inherent participant, we mean one which the verb semantically demands, rather than one which its meaning “metaphysically” demands. To give an example, in English ‘know’ (in the sense of the French ‘savoir’, the Spanish ‘saber’, etc.) has two linguistically inherent participants, namely the knower and the fact known. It is arguably a metaphysical necessity that knowing also requires a source of the knowledge: did it come from perception, inference, speech, etc.? But some languages, English included among them, do not mark the verb ‘know’ as having to specify this information. Thus source may be a “metaphysically” inherent participant, but it isn’t a linguistically inherent one. Sometimes, a verb whose semantic frame specifies n linguistically inherent participants in non-subject positions can occur in a well-formed sentence that contains n-1 non-subjects. And yet, the sentence in question is well-formed.

This, we think, is what “omission of a complement” comes to: the lack of isomorphism between what the semantic frame specifies as necessary, for the verb, and what nonsubjects may appear overtly in well-formed sentences containing the verb in question.

Our question here is: What exactly is going on in such apparently puzzling cases, especially with regard to the syntax and the semantic of such sentences?

II. Fillmore’s Account In answering our question, we begin with Charles Fillmore’s influential account. As he stresses, one crucial element that needs to be accounted for is that, in English and many other languages, not all verbs are like ‘follow’, ‘promise’ and ‘leave’, allowing null complements. Some verbs seemingly require that, when they appear in a sentence, each participant listed in their semantic frame must be phonologically realized. Null complements are simply not allowed with these verbs. Recall ‘lock’, for example. Its semantic frame identifies two linguistically inherent participants, the agent who does the locking, and the thing locked. But on Fillmore’s account, whereas ‘Aryn followed’ is

fine, (4) is simply ungrammatical in English. Similarly, (5) and (6) are ill-formed:

4. *Catherine locked 5. *Rob guaranteed 6. *Steven vacated

These last examples serve to highlight another point stressed by Fillmore (1986):

verbs which have quite similar meanings can nevertheless differ as to whether they permit a null complement. ‘Promise’ permits them, but the semantically close ‘guarantee’ does not; ‘leave’ permits them, but its semantic cousin ‘vacate’ does not.1 Two further important details about Fillmore’s view. First, it seems that the default reading of a sentence with a null complement varies, as a function of the verb in question, between definite and indefinite. Fillmore (1986) describes the contrast as follows. With some verbs, the missing element must be retrieved from what is “given” in the context; the unspoken referent is, in these cases, one definite thing, and its precise identity must be recoverable from the speech context. Verbs which are read this way, Fillmore labels ‘definite null complements’ (DNC). In contrast, there are verbs such that when the non-subject is omitted, the identity of the “missing element” is unknown or a matter of indifference. The unnamed thing is indefinite, if you will. Verbs which are read in the second way, Fillmore labels ‘indefinite null complements’ (INC). Examples will help clarify this contrast. In (7) and (8), the default reading is that a particular contextually salient thing was found out/followed.

7. Samir found out at midnight

8. Hershad followed quietly In (9) and (10), by contrast, something gets eaten/sung at midnight – but the precise

identity of the element eaten/sung is not at issue:

9. Omar eats at midnight

10. Khalid sings at midnight Fillmore (1986) offers a useful pragmatic heuristic for distinguishing DNC from INC verbs. With the former, it is odd for a speaker to say “A v-ed. I wonder what A v-ed?” Thus ‘Samir found out at midnight. I wonder what he found out?’ is a distinctly peculiar sequence for one person to say. In contrast, it is not pragmatically odd, in the case of an INC verb, to say the sequence “A v-ed. I wonder what A v-ed?” Notice, for instance, that someone who utters (10) can follow by saying ‘I wonder what he sings’ without any awkwardness. We would add, as a supplemental test, that in the case of DNC verbs “A ved” means something quite like “A v-ed it”. Whereas, in the case of INC verbs, “A v-ed” means something more like “A v-ed something” or “A v-ed stuff”. Thus the natural reading of (8) is close to ‘Hershad followed it quietly’, while the natural reading of (10) is closer to ‘Khalid sings something at midnight’. (We will have much more to say below, about what the precise truth-conditional contrast is here. But at present we are explaining

Fillmore’s view as he himself presents it.) We thus arrive at the following taxonomy:

Figure 1:

–  –  –

The last detail of Fillmore’s view has to do with what thing it is that permits, or does not permit, a null complement. We spoke above of “the verb” allowing/disallowing omission. But, for Fillmore, it seems rather to be a verb-on-a-sense that permits/prohibits null complements.2 Thus consider the contrasts below, lifted from Fillmore (1986: 101Apply’ a. I applied for the job → I applied b. They applied the bandage → *They applied 12. ‘Arrive’ a. She arrived at the summit → She arrived b. She arrived at the answer → *She arrived 13. ‘Hear’ a. I heard that you resigned → I heard b. I heard the song → *I heard 14. ‘Know’ a. They know that she resigned → They know b. They know her → *They know 15. ‘Leave’ a. She left home → She left b. She left this package → *She left 16. ‘Volunteer’ a. He volunteered to help you → He volunteered b. He volunteered his sons → *He volunteered In sum, for Fillmore the phenomenon of null complements has four sub-parts: a) some verbs which have n linguistically inherent participants listed in their semantic frame can nevertheless appear in fully grammatical sentences that do not contain n phonological realizers for those participants; b) some verbs allow this freely, others do not; c) when a phonological realizer is absent, at least two different kinds of default readings can be found, definite and indefinite, depending upon which verb is in play; d) it is not “the verb” so much as “the verb-on-a-sense” that seems to permit/prohibit null complements.

There are, we think, two key problems with the third plank of Fillmore’s view.

First, there is a problem (or rather, a cluster of them) with Fillmore’s description of the “meaning effects” of complement omission, several of them noted by Groefsema (1995).

Second, there is a problem with the very taxonomy that Fillmore offers, precisely because it posits two sub-varieties of null complements.

III. Problems with Fillmore’s Account of the “Meaning Effects” and Our Alternative We begin with issues about the supposed “meaning effects” of omission, on Fillmore’s view. First, on what “(in)definiteness” amounts to. Fillmore makes the unfortunate remark that that being marked indefinite means that the null complement “is obligatorily disjoint in reference with anything saliently present in the pragmatic context” (1986: 97).

We disagree. To begin with, ‘disjoint’ is surely too strong. As Groefsema (1995: 142)

notes, (17) is perfectly consistent with Ann eating some of the sandwiches:

17. John brought the sandwiches and Ann ate Here, the null complement of ‘ate’ is not disjoint with the salient sandwiches – though it is not wholly co-referential with it either. Second, and speaking of “reference”, we think that a better characterization of indefiniteness is what Fillmore hints at elsewhere in the paper: the null complement is interpreted as some kind of existential quantification, hence not as reference at all. The contrast is thus not “reference to something salient, versus reference to a particular but unknown object”. At best, it is reference versus (existential) quantification.

But even this isn’t quite right: in general, INC verbs are not subject to the sort of scope ambiguities one would expect, if they were covertly existentially quantified.

Notice, first, that (18) and (19) do not share the same readings:

18. No one ate something

19. No one ate Whereas (18) has a reading with ‘something’ given wide scope – there is a dish, say the spinach spaghetti with tofu balls, that no one ate – (19) has no such reading. Compare

also (20) with (21), and (22) with (23):

20. Joaquim eats everyday

21. Joaquim eats something every day

22. Joaquim ate all day

23. Joaquim ate something all day If we must find a synonym for (19) that has something overt in the complement spot, ‘No one ate stuff’ would serve better. In light of these sentences, Fillmore will need to say that INC verbs are read as “V-ed stuff”.

Fillmore’s gloss of definiteness is incorrect too. In particular, it is not the case that the null complement in DNC cases must refer to some particular entity made salient.

Instead, the null complement can function as a variable bound by a higher quantifier.

Thus, in a situation in which various wives are hiding various different secrets, (24) can be read as “Each husband found out his wife’s secret”. Similarly, the second sentence in (25) need not mean that no one left the salient place at which the sentence is spoken, nor any other one contextually spot. Rather, each departure-candidate can be said to have not left whichever threatened theme park s/he was at. Finally, what the press follows in (26) is not some salient individual from the context; rather, the press follows each respective important politician.3

24. Each husband found out

25. There were terrorist threats to several of Disney’s theme parks. But not one customer left

26. Whenever an important politician takes a trip to Moscow, the press follows A better description of DNCs, then, would be that their null complements behave like variables: they may be left free, in which case they must pick out an object salient in the context; or they may be bound by a quantifier (supplied explicitly, higher in the sentence, or by other means).

Notice that this latter, alternative, proposal explains another bit of data which bedevils Fillmore’s distinction. As he describes it, one would expect each verb sense to be either [+ definite] or [- definite]. The twain should never meet, and there should be no cases which are hard to classify. But some verbs do seem hard to classify, and do seem to permit either reading – with pragmatic context being crucial. Consider: Is ‘aim’ INC or DNC? One can imagine the null complement referring to some salient thing, as in (27), but one can also imagine a use of ‘aim’ without a phonologically realized object, that is more existential in tone, as in (28).4

27. Jean Pierre saw the deer, and aimed carefully

28. To avoid accidents, it’s important to aim carefully Similar points apply to ‘disagree’, and a host of other cases. Crucially, if null complements are read as variables, as we propose, then it’s no surprise that they may be read as free, or as quantificationally (including existentially) bound, depending upon the context.

In sum, regarding INC verbs, at a minimum one must excise talk of “disjoint reference”, and substitute the idea that INC verbs are read as “V-ed stuff”. Regarding DNC, rather than taking such verbs to be always referential (to some contextually salient something), we would urge that they are read as if they had a variable -- which is available for binding, or may be read as deictic. But, for reasons that will emerge immediately below, even this isn’t enough to fix the overall account.

Pages:   || 2 | 3 |

Similar works:

«Отзыв официального оппонента по диссертации Панасенко Людмилы Александровны на тему «Интерпретирующий потенциал лексических категорий» на соискание ученой степени доктора филологических наук по специальностям: 10.02.04 Германские языки, 10.02.19 Теория языка Лейтмотивом...»

«Improving Intrusion Detection Performance Using Keyword Selection and Neural Networks Richard P. Lippmann Robert K. Cunningham MIT Lincoln Laboratory, Rm S4-121 MIT Lincoln Laboratory, Rm S4-129 244 Wood Street 244 Wood Street Lexington, MA 02173-0073 Lexington, MA 02173-0073 rpl@sst.ll.mit.edu rkc@sst.ll.mit.edu phone: (781) 981-2711 Abstract The most common computer intrusion detection systems detect signatures of known attacks by searching for attack-specific keywords in network traffic....»

«MAC 2312: Calculus II Online Version Fall 2015 CONTACT INFORMATION: COURSE COORDINATOR and LECTURER: KL Chui, LIT 376 Office Phone & Hours: 352-299-2244, hours by appointment Email: Use Canvas mail by clicking on Inbox in the upper-right-hand corner. TEACHING ASSISTANT for MAC2312: Jason Nowell, LIT 475 Office Phone & Hours: 352-294-2378, hours TBD Email: Use Canvas mail by clicking on Inbox in the upper-right-hand corner. TEXT: Calculus, Early Transcendentals by Rogawski, 2e...»

«Successful Interviewing An interview is a meeting between an employer and an applicant to discuss a job. While job interviewing for most people may not seem to come naturally, there is much that can be done in advance to increase your effectiveness as a candidate. Some individuals believe that just being themselves is sufficient for successful job interviewing. However, you are participating in a highly competitive selection process. You need to know how to effectively sell yourself,...»

«Stormwater Control Devices Filtration Basin A SHALLOW BASIN WITH ENGINEERED OR AMENDED SOIL AND AN UNDERDRAIN SYSTEM Filtration basins function by detaining stormwater in the basin. As stormwater infiltrates through the amended soil, sand, or engineered media, pollutants are filtered and adsorbed onto soil particles. Treated stormwater is directed to the receiving stream via the underdrain system. Filtration basins may be shaped like ponds or channels. To improve pollutant removal, the basin...»

«Bachelor Thesis Impact of Future E-Mobility Fleets on suburban Medium-Voltage Grids By Guillermo Moraleda angefertigt am Karlsruher Institut für Technologie (KIT) Institut für Elektroenergiesysteme und Hochspannungstechnik Institutsleiter: Prof. Dr.-Ing. Thomas Liebfried Betreuer: Dipl.-Ing. Simon Wenig April 2015 AUTORIZACIÓN PARA LA DIGITALIZACIÓN, DEPÓSITO Y DIVULGACIÓN EN ACCESO ABIERTO ( RESTRINGIDO) DE DOCUMENTACIÓN 1º. Declaración de la autoría y acreditación de la misma. El...»

«Contemporary Music Review Vol. 25, No. 1/2, February/April 2006, pp. 69 – 79 Embodied Sound: Aural Architectures and the Body Gascia Ouzounian This article examines two sound installations distributed on CD: Maryanne Amacher’s Sound Characters (Making the Third Ear) (1999) and Bernhard Leitner’s ¨ KOPFRAUME (HEADSCAPES) (2003). The author undertakes an embodied reception of these works, experimenting with new models of listening and analysis that take into consideration aspects of the...»

«VOLUME 76, NUMBER 41 Longview Pioneer Lions Club rd In our 83 year 1923-2006 Home club of Past Int’l Dir. – The Late Arne Peterson Past Dist. Govs – Frank Brown, The Late Tom Hinshaw, Mike Parker, Sam Price, Jim Reinecke, The Late Stuart Wasser and Hal Palmer. Walt Specht, 2004-2005 Lion of the Year Kim Pacsuta, 2004-2005 Unsung Lion of the Year Deena Martinsen, 2004-2005 Rookie of the Year John Claypool, Editor Bethanie Young, Publisher April 6, 2006 Bill LaVigne, Web Master...»

«NOT MEASUREMENT SENSITIVE MIL-HDBK-17-2F Volume 2 of 5 17 JUNE 2002 SUPERSEDING MIL-HDBK-17-2E Volume 2 of 5 24 MAY 1999 DEPARTMENT OF DEFENSE HANDBOOK COMPOSITE MATERIALS HANDBOOK VOLUME 2. POLYMER MATRIX COMPOSITES MATERIALS PROPERTIES This handbook is for guidance only. Do not cite this document as a requirement. AMSC N/A AREA CMPS DISTRIBUTION STATEMENT A. Approved for public release; distribution is unlimited. MIL-HDBK-17-2F Volume 2, Foreword / Table of Contents FOREWORD 1. This Composite...»

«About the book. The House of Mirth, published in 1905, marks Edith Wharton's emergence as one of America's greatest writers. Although Wharton had previously published two collections of stories, The Greater Inclination (1899) and Crucial Incidents (1901) and the novel The Valley of Decision (1902), her decision to write about fashionable New York, a world she had been steeped in since infancy, brought her immediate success and recognition. As she wrote in her autobiography, A Backward Glance...»

«Quartz Plate Calorimeter as SLHC Upgrade to CMS Hadronic EndCap Calorimeters U. Akguna, P. Brueckena, K. Cankocaka, L. Cremaldib, F.Durua, I. Dumanoglud, J. Merloa, A. Penzoe, Y. Onela, D. Winnc a University of Iowa, Iowa City, IA 52242, USA b University of Mississippi, University, MS 38677, USA c Fairfield University, Fairfield, CT 06430, USA d Cukurova University, Adana, Turkey e University of Trieste and INFN, Trieste, Italy (for CMS Collaboration) Abstract. Due to an expected increase in...»

«DECEMBER, 19x9. [No. 349. VOL. XLVII.] UK THE MANCHESTER GRAMMAR 5CHGDL MAGAZINE) OFFICIAL ORGAN O T E O D FH L MANCUNIANS ASSOCIATION f PRINCIPAL CONTENTS: Occasional Notes. O.M.A. Notes and O.M. Chronicle. Dr. Mumford's Book on the Grammar School. Le Mot Juste. Sports, Societies, etc. A P U B L I S H E D B Y J G A I T &CO.,LTD. JOHN D A L T O N J T M A M C H E J T £ R ^ P R I N T E D B Y H R A W O N & C O, 1 6 N E W B E O W M ST M A M C l i l i i T E R Annual Subscription to the O.M.A....»

<<  HOME   |    CONTACTS
2016 www.dis.xlibx.info - Thesis, dissertations, books

Materials of this site are available for review, all rights belong to their respective owners.
If you do not agree with the fact that your material is placed on this site, please, email us, we will within 1-2 business days delete him.