The morphosyntax of Participles and Nominalizations in Swedish. Bj¨orn Lundquist October 7, 2009
1
Participles, nominalizations and other syntactic categories • Participle: A verb dressed up as an adjective - An adjective derived from a verb - Semantically, syntactically and morphologically somewhere in between an adjective and a verb. • (de-verbal) Nominalization: A verb dressed up as a noun - a noun derived from a verb Semantically, syntactically and morphologically somewhere in between a noun and a verb.
Participles can be more or less verb-like: (1)
a. b.
F¨oretaget erbj¨od dem nya jobb The company offered them new jobs De blev erbjudna nya av f¨oretaget They were offered new jobs by the company
The participle in (1-b): • denotes an event, assigns structural case to a complement, and has the same argument structure as the corresponding verb (verbal properties). • carries number/gender agreement morphology (though the gender is not visible in the plural), does not carry tense marking and appears in the complement of copula (adjectival properties). (2)
a. b.
Han st¨angde d¨orren He closed the door D¨orren a¨r fortfarande st¨angd. The door is still closed
(2-b) differ in both event structure and argument structure from the corresponding verb (2-a). In short, it carries none of the morpho-syntactic or semantic characteristics of a verb.
1
2
Non-finite verbs
Nominalizations can be more or less verb-like: As an example, the nominalizing -ing-suffix in English can form nominals from verbs, and the nominals can carry more or less semantic, syntactic and morphological similarities with the corresponding verb. (see .e.g. Lees 1964, Chomsky 1970, Ross 1973, Abney 1987 and Grimshaw 1990 for discussion of different types of ing-nominals). (3)
He paints pictures featuring the recent disturbances in Los Angeles.
In (4), three different types of nominalization of the same verb (paint) are given, all derived with the suffix -ingThe three examples show a declining amount of verbal properties: (4)
a. b. c.
[John’s painting a picture featuring the recent disturbances in Los Angeles] caused a huge riot among the art people. A classic example is [John L’s painting of a picture featuring the recent disturbances in Los Angeles]. [A painting (*of a picture featuring the recent disturbances in Los Angeles) by John L] hung on the wall
(4-a) (so-called POSS-ing): it assigns accusative case to its internal argument, and it denotes an event (verbal properties), BUT no Tense marking, Possessive subject and occupies argument position (nominal properties) (4-b) (so-called ing of -nominalization or ‘mixed’ nominalization): Denotes an event, takes an internal argument (verbal properties) BUT no tense, no accusative case, occupies argument position etc. (nominal properties) (4-c) (result nominal): Denotes an object, no internal argument (only nominal properties) Why study participles and nominalizations? 1. To get a better understanding of lexical/syntactic categories. 2. To get a better understanding of the fine-grained semantic and morphosyntactic structure of verbs. 3. To get a better understanding of syncretism. 4. To get a better understanding of the syntax-lexicon interface. 5. To get a better understanding of what seems to be semi-transparent/semi-productive morphology (i.e., derivational morphology).
2
Participles and nominalizations in Swedish
The scope of the Thesis: Two types of participles + two types of nominalizations in Swedish. As shown below, it seems like each of them can fill many functions, and to carry more or less similarities with verbs (just like ing in English):
Non-finite verbs
3
• Passive (past) participle (= -de-participle) (5)
a. b. c. d.
de nyligen planterade blommorna (attributive eventive participle) the recently planted flowers Mannen blev erbjuden ett nytt jobb. (predicative eventive participle) The man was offered a new job D¨orren a¨r fortfarande st¨angd. (Stative participle) The door is still closed Jag a¨r v¨aldigt imponerad. (gradable stative particple) I am very impressed
• Present participle (= -nde-participle) (6)
a. b. c. d. e.
den springande mannen (attributive eventive participle) The running man Han kom springande. (eventive participle) He came running Boken ¨ar v¨aldigt fascinerande. (adjectival (gradable) participle) The book is very fascinating en analys r¨orande de ekonomiska skillnaderna (“prepositional” participle) an analysis concerning the economical differences med sockerf¨orgiftningen spridande sig i min kropp.(“big” gerundive) with sugar.poisoning.def spread.nde REFL in my body ‘With the sugarpoisoning spreding in my body’
• Nominalizations 1 ((n)ing-nominalization) (7)
a. b. c.
Napoleons avrttning av hertigen av Enghien (Event nominal) Napoleons execution of the Duke of Enghein Efter simningen slog jag av TV’n. (Simple event nominal/event naming) After the swimming I turned off the TV en blandning av nytt och gammalt (Result nominal) a mix of old and new
• Nominalization2 (nde-nominalization) (8)
a. b.
hans st¨andiga tjatande om ov¨asentliga ting (event nominal) his constant nagging about unimportant things meddelandet n˚ adde inte fram The message didn’t arrive (Result?)
1. Claim 1: The three (or four, depending on how you count) are productively used. 2. Claim 2: (Part of) the meaning of the verb will transparently be present in the derived participle/nominalization.
4
Non-finite verbs
Reasonable assumption: The verb plus the ending is not stored in the lexicon as a simplex item. In other words, the lexical entries (9) (b-d) do not exist: (9)
a. b. c. d.
samlaV = collectV samlandeA,N = collectingA,N samlingN = collectionN , collectingN samladA = collectedA
Rather, we should have lexical entries like the following: (10)
a. b. c. d. e.
samlaV = collectV -nde(1)adj (-verb) -deadj (-verb) -nde(2)adj (-verb) -(n)ingnom (-verb)
(meaning: -nde(1) and -de are functions from verbs to adjectives, -(n)ing and -nde(2) are functions from verbs to nouns. However, the entries in (10) can not be completely correct either. Why? Three reasons:
2.1 All of the suffixes in (10) select items that belong to the category V, but none of them can attach to all items of the category V. • Some restrictions on (n)ing: Can not combine with unaccusative verbs (11-a), stative verbs (11-b) + a lot of ‘idiosyncratic’ exceptions (11-c): (11)
a. *Artistens pl¨otsliga f¨orsvinning f¨orbryllade fansen artist.def.poss sudden disappear.(n)ing confused fans.def.pl *The artist’s sudden disappearing confused the family b. *hans hatning av grannen his hate.(n)ing of the neighbor. def c. efter ??dansningen/vandringen var han tr¨ott after dance.(n)ing.def/hike.(n)ing.def was he tired “After the dancing/hiking he was tired”
• Some restrictions on -ndeN om : Cannot combine with e.g. two types of stative verbs: (i) spatial/prepositional verbs (12-a), and (ii) object oriented psych-verbs (12-b): (12)
a. *kakans inneh˚ allande av n¨otter cookies contain.nde of nuts b. *jordb¨avningens skr¨ammande av barnen earthquakes frightening of the children
• Some restrictions on -de: Cannot combine with (i) a fairly well-defined group subtype of monotransitive verbs (basially ‘unergative’ verbs) (13-a) and (ii) a slightly less well-defined
Non-finite verbs
5
group of transitive verbs (mainly a sub-group of the stative verbs, namely the ‘prepositional’ verbs) (13)
a. *den skrattade mannen the laughed man b. *n¨otter ¨ar inneh˚ allna i kakan Nuts are contained in the cookie
• Some restrictions on -ndeadj : Cannot combine with non-transitive verbs (i.e., weather verbs) (14-a) and verbs that lack an infinitival form (14-b): (14)
a. *det blev regnande hela dagen it bli.past rain. nde whole day int. ‘it kept/ended up raining the whole day’ b. *m˚ astande musting
2.2 Differences in interpretation of the derived forms. • Participles (15)
a.
de anl¨andande/anl¨anda g¨asterna the arriving/arrived guests
In this case, the nde-participle refers to an ongoing event and the de-participle refers to completed event (though, for other verbs, there will be other differences). • Nominalizations (16)
a. b.
En blandning av vatten och jord t¨ackte golvet a mix of water and dirt covered the floor. Jag tar hand om blandandet av drinkar I take care of the mixing of drinks.
The (n)ing-nominal in (16-a) denotes the result of a mixing event, while the nde-nominal in (16-b) denotes the very event of mixing.
2.3 The derived forms under discussion show certain quirks in their distribution, which indicates that the syntax does not treat these elements as simple, atomic adjectives and nouns (the discussion below will only deal with participles, see e.g. Grimshaw (1990) for discussion about restricted distribution of different types of de-verbal nouns):
6
Non-finite verbs
• Most present participles (-nde-participles) cannot occur in the complement of the copula vara (17-a), and can further not be graded (17-b): (17)
a. *han var skrattande He was laughing b. den (??mest) skrattande mannen the (??most) laughing man
• Many passive (past) participles (de-participle) formed from intransitive verbs cannot occur in the complement of the copulas bli and vara: (18)
Olyckan ??¨ar/*blev intr¨affad ig˚ ar accident.def vara.pres/bli.past yesterday “*The accident is/was occurred yesterday
We end up with three important questions: 1. How to capture the range of meanings that each suffix can give rise to? (i.e., the issue of ‘syncretism’) 2. How to capture the differences between the four suffixes? (Most importantly: how to capture the differences between the -nde-nominalizations and the (n)ing-nominalizations, and how to capture the differences between the de-participles and nde-participles). 3. How to capture gaps in (i) distribution and (ii) selection. But first to some notes on the theory.
3
The model: what and why
Nano-syntax, as developed at the University of Tromsø during the last years. 1. Late insertion: The syntax operates on abstract morpho-syntactic, semantically motivated features, like plural, Init and Place (see e.g. Ramchand 2008 and Svenonius 2006). A syntactic representation is built up by merging these features in an order following an underlying “functional sequence”. Once the hierarchical structure is built up, lexical insertion takes place, whereby the abstract features get replaced by lexical items, that carry matching features. 2. Hierarchical structure all the way down: Word syntax/Morphology and Phrasal syntax are handled in the same module. Note: My dissertation ignores many important issues that deal with phrasal syntax (mainly placement/introduction of DP arguments of verbs). I do however argue strongly against the “lexicalist hypothesis” as proposed in Chomsky (1972), and developed in Wasow (1977), according to which certain morphology is introduced a pre-syntactic lexical module (derivational morphology), whereas others are introduced in the syntax.
Non-finite verbs
7
3. Feature specified “roots”/no strict distinction between lexical and functional items: For the this discussion, Verbs will be the only relevant domain. Lexical entries for verbs will look like below (from Ramchand 2008). Verbal entries carry information about sub-eventual structure and the co-reference relations between the arguments: (19)
a. b. c. d.
samlaT ransitive (‘collect’): [Initi Proci , Resj ] f¨orsvinnaU nacc (‘disappear’): [Initi , Proci , Resi ] skrattaU nerg. (‘laugh’): [Initi Proci ] chaseT rans (‘chase’): [Initi , Procj ]
4. Spell out of non-terminals: One and the same lexical item can contain more than one feature. For example, Ramchand gives the verb destroy the feature specification [init], [proc], [res]. This means that one and the same root can lexicalize many terminal nodes, or a full sub-tree. This can either be captured by the means of head-movement, or by allowing spellout of non-terminal nodes. I will go for the latter choice here, following recent work by Michal Starke, presented in various seminars (see also Caha 2007 for more details and arguments why spell out of non-terminals is necessary). 5. “Overspecification”/Superset Principle: A system that allows spell out of non-terminals needs the exact opposite to the subset principle, namely the superset principle, as has been worked out by Michal Starke. Caha (2007) gives the following definition of the superset principle (observe that it is formulated to mirror Halle’s definition of the subset principle) • The Superset principle: The phonological exponent of a Vocabulary item is inserted into a node if the item matches all or a superset of the grammatical features specified in the node. Insertion does not take place if the Vocabulary item does not contain all features present in the node. Where several Vocabulary items meet the conditions for insertion, the item containing less features unspecified in the terminal morpheme must be chosen. Relevance for verbal items: (20)
SamlaT ransitive (‘collect’): [Initi Procj , Resj ]
Can be inserted in at least the following syntactic structures: (21)
a. b. c.
[ Resi ] [ Proci , Resi ] [Initi Procj , Resj ]
But not in the following: (22)
[Initi , Proci , Resi ]
On the other hand, the unaccusative verb f¨orsvinna can be inserted in (22), as well as in (21-a) and (21-b), though not (21-c).
8
Non-finite verbs
(23)
skrattaU nerg. (‘laugh’): [Initi Proci ]
can be inserted in the following contexts (at least): (24)
a. b.
[Proci ] [Initi Proci ]
but not in any of the following contexts: (25)
4
a. b.
[Initj , Proci ] [Initi , Proci , Resi ] (unless an independent item lexicalizes Res)
Back to the three questions!
4.1
Polysemy/syncretism
Abney (1987): a syntactic account of nominalizations in English: a morpheme can attach at different heights in the syntactic structure - combined with a decomposed verb phrase: Eventive (n)ing-nominal (26)
Nom- [Init [ Proc [ Res]]]
(27)
a. b.
H¨angningen av Saddam Hussein The Hanging of Saddam Hussein h¨angning - (n)ing- [Initi [ Procj [ Resj ]]] = “non-reflexive” interpretation
Result nominalization (28)
Nom- [Res]
(29)
a.
b.
En tidig Beatles-inspelning hittades p˚ a vinden an early beatles-in.play.ning found.past.pass on attic.def ‘An early Beatles-recording was found in the attic.’ Inspelning - (n)ing- [Res] (note, possible “reflexive” interpretation)
Eventive passive (30)
Ptcp- [Init [Proc [Res ]]]
(31)
a. b.
Han ¨ar/blev h¨angd He was/is hanged h¨angd: de- [initi Procj [Resj ]] (note, no reflexive interpretation)
Stative Passive: (32)
Ptcp- [Res]
(33)
a. b.
D¨orren a¨r fortfarande st¨angd The door is still closed st¨angd: de - [Resi ] (note, possible reflexive/anti-causative interpretation)
Non-finite verbs
4.2
9
Differences between the suffixes
Solution: The morphological structure of the forms in question (chapter 2): 1. -(n)ing: Attaches before Lexical Class markers/theme vowels (i.e. first conjugation -a), but after verbalizer er: (34) (34) (34) (34)
a. recens-er-a - recenser-ing (*recens-er-a-(n)ing) (*recens-(n)ing) simm-a - sim-ning (conj. 1) l¨op - l¨op-ning (conj. 2) sy - upp-sy-ning (conj. 3)
2. Passive participle -de (and other typical inflectional suffixes): Attaches outside theme vowel/lexical class marker. (I’m leaving strong and irregular verb outside the discussion for now): (35)
a. b. c.
tv¨att-a - tv¨att-a-d (conj. 1) bygg - bygg-d (conj. 2) (be)bo - (be)bo-dd (conj. 3)
3. Present participles/nominalizing (-nde: Attaches outside the infinitival -a: (36)
a. b. c.
recens-er-a - recens-er-a-(a)-nde (conj. 1) bygg - bygg-a-nde (conj. 2) bo - bo-ende (the /e/ before -nde is presumably epenthetic.)
• The present participle/nde-nominal contains infinitival /a/ and inherits the syntactic/semantic properties of the infinitival. • The infinitival in Swedish is never event structurally or argument structurally “reduced” - always selects for InitP • -de and (n)ing attach to something plausible smaller - i.e. something that does not need to be a full InitP (i.e., Proc or Res)
Plausible hypothesis for the role of the infinitival /a/: (37)
Generalization Concerning Infinitival /a/ The infinitive /a/ maps a complete argument structure to case-structure/functional structure. More specifically, it creates a position for an external argument/subject.
As discussed in chapter 3: 1. The nde-forms are never event-structurally reduced.
10
Non-finite verbs
2. The nde-forms are never argument-structurally reduced. 3. The nde-forms can take on vP(i.e. InitP)-external aspect values (especially, iterative). 4. The de-participles and (n)ing-nominals can be event-structurally reduced. 5. The de-participles and (n)ing-nominals can be argument-structurally reduced. 6. The de-participles and (n)ing-nominals cannot take on vP (i.e. InitP) external aspect values.
4.2.1
The alternative - Aspect
Several proposals out there that connect Swedish ande/English ing with progressive/imperfective. (Loman 1964 and Thur´en 2007 for Swedish, e.g. Pustejovsky 1991. Several proposals connect passive participle marking with perfective (Embick 2004 for English). Closer look at the data (chap 3) Aktionsart/Lexical aspect survives in all the relevant forms, i.e, the aspect hypothesis must be wrong.: 1. Present participles can get “perfective” interpretations, if the verb is telic (this is only easily seen for a subgroup of achievements). 2. Passive participles can get “imperfective” interpretations, if the the verb is atelic (i.e. State or Activity). 3. -nde-nominalizations can get a perfective interpretation, if the verb is telic. 4. (n)ing-nominals can be either imperfective or perfective. • How can I capture aspectual differences? (38)
De anl¨andande/anl¨anda g¨asterna the arriving/arrived guests
Why no imperfective reading of the passive participle: Passive cannot contain outer aspectual features. Why no perfective reading of the present participle: Use simplest form possible! This should follow from the elsewhere condition as stated in the superset principle.
4.3
Lexical entries for the suffixes
• The four (or three) different suffixes differ only in attachment height and value of Gender feature: (39)
a.
Interpretable gender: referential index (Baker 2003)
Non-finite verbs
11
b.
Uninterpretable gender: “slot” for gender agreement - necessary for combining with nouns in certain contexts.
a.
(underived) Nouns have gender features in their lexical entries (just like verbs have sub-eventual features) (underived) Adjectives have a gender “slot”/uninterpretable gender in their lexical entry.
(40)
b.
Lexical entries for the suffixes in question: (41)
5
a. b. c. d.
-ndenom : Interpretable Gender features, outside infinitival. -a -ndeP art : Uninterpretable Gender features, outside infinitival -a -(n)ing: Interpretable Gender features, inside infinitival -a -de: Uninterpretable Gender features, inside infinitival -a
Part 2: the suffixes one by one
The system as set up so far overgenerates: Predictions: 1. All verbs should have at least one (n)ing-nominal, and at least one de-participle (most verbs should have two or three of these). 2. All verbs should have a nde-participle and a -nde-nominalization. 3. The participles should have the same distribution as regular, underived adjectives (and the nominalizations should have the same distribution as underived nouns). None of the predictions above fall out. The second part of the dissertation aims at figuring out why. Two main strategies for dealing with the “gaps”: 1. Spell-out related explanations: • A: “Blocking” on the “word”-level: A morphologically less complex item is inserted instead of the overtly derived form. (Explains some of exceptions to point 1 above). • B: “Blocking” on the phrasal level: A morphologically less complex item is inserted instead of e.g. a copula + a participle. (Explains some of the exceptions to point 3 above). 2. The mysterious word-class factor: deeper, conceptual restrictions on noun- and adjective formation.
5.1
Spell-out related factor 1
Word-level blocking (see e.g. Embick and Marantz 2008 for discussion):
12
Non-finite verbs
5.1.1
(n)ing
Some verbal entries have gender/declension class information stored with them in addition to event/argument structure features: Complex events: Gen [ Proc [ Res ]] (Init is possibly present) (42)
a.
b. (43)
a. b.
k¨op/*k¨opning av hus buyN of house ‘House-buying’ s¨aljning/*s¨alj av hus ‘selling of house’ k¨op: [Init, Proc, Res, Neut] s¨alj: [Init, Proc, Res ] (in nominal contexts, (n)ing kicks in as a default gender)
Simplex events/processes: Gen [ Proc ] (44)
a. b.
(45)
a. b.
Under dansen/??dansningen b¨orjade jag bli tr¨ott. ‘during the dans/dance-ing I started to get tired’ Under vandringen/*vandren b¨orjade jag bli tr¨ott. ‘During the hiking/hike I started to get tired’ dansa: [Init, Proc, Common Gen] vandra: [Init, Proc ] (in nominal contexts, (n)ing kicks in as a default gender)
Result nominals: Gen [ Proc ] (46)
a.
b.
(47)
a. b.
Det st˚ ar en h¨og stapel/#stapling med gamla l˚ ador p˚ a golvet. it stand a tall pile/pile-ing with old boxes on floor.def ‘There’s a big pile of old boxes on the floor.’ Han har en stor samling/*sam(m)el med gamla serietidnignar. he has a big collect(n)ing/collect with old cartoons ‘He has a big collection of old cartoons’ Stapla: [Init, Proc, Res, Common gen ] samla: [init, Proc, Res ] (in nominal contexts, (n)ing kicks in as a default gender)
Gender feature is always related to the smallest possible nominalization! (You need to specify which sub-event the gender feature is associated with! ) (48)
en maskin f¨or stapling/*stapel av l˚ ador a machine for piling/*pile of boxes
5.1.2
-de/Passive Participles
(49)
a. b.
den nyligen st¨angdapart /¨oppnadepart d¨orren The recently closedpart /openedpart door D¨orren a¨r forfarande st¨angdpart /¨oppenadj The door is still closedpart /openadj
A verb like ankomma (‘arrive’) presumably has a Res-projection, but no stative participle:
Non-finite verbs
(50)
13
De (*fortfaranda) ankomna g¨asterna The (*still) arrived guests
Why? The result predicate corresponds to a spatial predicate like ‘Here’ - the stative participle is blocked by ‘here’ which is presumably lighter feature-wise. Same think can be said for e.g. double object verbs: Res = Poss, and the feature Poss can be spelled out by items that are lighter than a passive double object verb (e.g. the possessive ’s, or preposition like to or of). Important note: no underived adjectives realize a complex event structure (i.e., No underived adjectives can realize the following structures [Proc [ Res ]] or [Init [ Proc [ Res ]]]) 5.1.3
-nde
Cases of blocking are harder to find, as expected, since lexical items in Swedish do not by themselves realize a complex VP-structure + the function spelled out by the infinitival a. Possible exceptions: Verbs that take clausal complements: (51)
a. b.
(52)
a. b.
5.2
Han p˚ astod att jorden a¨r platt He claimed that world.def is flat p˚ ast˚ aendet att jorden a¨r platt claim.nde.def that world.def is flat ‘The claim that the world is flat’ han fr˚ agade om jorden ¨ar platt he asked if the world is flat fr˚ agan/*fr˚ agandet om jorden a¨r platt ask.def/asknde.def if the world is flat ‘The question if the world is flat’
Spell-out related factor 2
‘Blocking’ that takes place above the word level. In a framework that employs late insertion + spell out of non-terminals (or simply a system where one item can spell out many functional heads). This discussion only deals with the participles, and gives a suggestion on how to deal with the deficient distribution of certain participles: 5.2.1
Present participles/nde-participle
Eventive present participles can be used attributively (and in the complement of certain light verbs), but not in complement of the copula vara, unlike other adjectives: (53)
a.
den skrattande mannen the laughing man (i) *han var skrattande He was laughing
Blocking by the simple present tense form:
14
Non-finite verbs
(54)
Han skrattar He laughs
The present participle lacks one or two features that the present tense has: (1) Tense/finiteness, and (possibly) (2) a function that licenses a predicational relationship between the verb-phrase and the subject. The copula can carry both these features, and so can the present tense. The present tense is chosen over the complex form. (This should either follow from the elsewhere condition of the superset principle, though it is possible that one has to postulate another rule that says that simple forms are preferred over complex forms). 5.2.2
Passive past participle
Event denoting passive participles formed from unaccusative verbs are blocked in the complement of the copula vara, unlike eventive and stative passive participles formed from transitive verbs, stative participles formed form unaccusative verbs and underived adjectives (but just like eventive present participles). In other words, for passive participles the blocking only kicks in when the participle plus the copula vara would spell out the exactly the same structure as the simplex verb. (55)
5.3
a.
den nyligen intr¨affade olyckan The recently occurred accident b. *Olyckan var/blev intr¨affad ig˚ ar accident.def vara.past/bli.past yesterday ‘*The accident was occurred yesterday’
The mysterious word class factor
Not all syntactic structures can be nominalized! Not all syntactic structures can be turned into participles! Clear cases: 1. Most unaccusatives cannot be nominalized until they are as big as an infintival (i.e., when the full vP is built up - i.e., they have only -nde-nominals, and no (n)ing- or zero derived nominals). 2. Most unergatives cannot be turned to adjectives until they are as big as an infinitival.(i.e., when the full vP is built up - i.e., they have only present participles and no de-participles or corresponding zero-derived nominals) Hale and Keyser (1993): Unergative verbs are denominal - unaccusative verbs are de-adjectival: (56)
a. b.
he danced (unergative - dancenoun ) the door opened (unaccusative - openadj )
• Hale and Keyser: Lexical categories are primitives - derive argument structure from lexical categories. • Nanosyntax: Lexical categories are not primitives - certain syntactic structures can surface in the shape of certain lexical/syntactic categories.
Non-finite verbs
15
Not only true for “small” (pre-infintival) structures, but also true for the nde-forms: TABLE 1: Predicative and non-predicative inf+-nde inf+-nde 1 Inf+-nde 2 Inf+-nde 3 Participle Verbal Adjectival Prepositional Nominal Event Nom. * * The verbs that form adjectival and prepositional present participles do not have corresponding nominals – the Hale and Keyser facts must be valid at bigger structures as well!
6
Conclusion 1. To understand nominalizations and participles you have to pay attention to a fine-grained syntactic decomposition. 2. The interpretational differences between different types of nominalizations and participles (derived with the same or different morphemes) mainly have their origin in the verbal root. 3. The study of participles and nominalizations opens up for a more detailed understanding of lexical categories, and specifically how they relate to argument structure and event structure.
References Abney, Steven. 1987. The English Noun Phrase in its Sentential Aspect. Ph.D. thesis, MIT, Cambridge, Ma. Baker, Mark C. 2003. Lexical Categories: Verbs, Nouns and Adjectives. No. 102 in Cambridge Studies in Lingusitics. Cambridge University Press, Cambridge. Caha, Pavel. 2007. The superset principle. Ms. University of Tromsø . Chomsky, Noam. 1970. Remarks on nominalization. In Readings in English Transformational Grammar , edited by Roderick A. Jacobs and Peter S. Rosenbaum, pp. 184–221. Georgetown University Press, Washington, D.C. Chomsky, Noam. 1972. Studies on semantics in generative grammar . Mouton, The Hague. Embick, David. 2004. Structure of the resultative participles in english. Linguistic Inquiry 35 3: 355–592. Embick, David and Alec Marantz. 2008. Architecture and blocking. Linguistic Inquiry 39 1: 1–53. Grimshaw, Jane. 1990. Argument Structure. Linguistic Inquiry Monographs; 18. MIT Press, Cambridge. Hale, Kenneth and Samuel Jay Keyser. 1993. On argument structure and the lexical expression of syntactic relations. In The View from Building 20: Essays in Linguistics in Honor of Sylvain Bromberger , edited by Kenneth Hale and Samuel Jay Keyser, no. 24 in Current Studies in Linguistics, pp. 53–109. MIT Press, Cambridge, Ma.
16
Non-finite verbs
Lees, Robert B. 1964. The grammar of English nominalizations. Mouton, The Hague. Loman, Bengt. 1964. Verbalsubstantiv p˚ a ning och -ande i nusvenskt riksspr˚ ak. In Nutidssvenska - Uppsatser i Grammatik , edited by Ture et al Johannison, pp. 45–90. Almqvist och Wiksell, Stockholm/G¨oteborg/Uppsala. Pustejovsky, James. 1991. The syntax of event structure. Cognition 41: 47–81. Ramchand, Gillian. 2008. Verb Meaning and the Lexicon: A First Phase Syntax . Cambridge University press, Cambridge. Ross, John Robert. 1973. Nouniness. In Three dimensions of linguistic research, edited by Osamu Fujimura, pp. 137–257. TEC Company ltd., Tokyo. Svenonius, Peter. 2006. Spatial p in English. In The Cartography of Syntactic Structures. Vol 6 , edited by Guglielmo Cinque and Luigi Rizzi. Oxford University Press, Oxford. Thur´en, Camilla. 2007. The Syntax of the Swedish Present Participle. Ph.D. thesis, Lund University, Lund. Wasow, Thomas. 1977. Transformations and the lexicon. In Formal Syntax , edited by Peter Culicover, Thomas Wasow, and Joan Bresnan, pp. 327–360. Academic Press, New York.
