Construction Morphology

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/220531577
Construction Morphology
Article in Language and Linguistics Compass · July 2010

DOI: 10.1111/j.1749-818X.2010.00213.x · Source: DBLP
CITATIONS READS
385 9,151
1 author:
Geert Booij
Leiden University
171 PUBLICATIONS 6,098 CITATIONS
SEE PROFILE
All content following this page was uploaded by Geert Booij on 29 November 2017.
The user has requested enhancement of the downloaded file.

Language and Linguistics Compass 3/1 (2010): 1–13, 10.1111/j.1749-818x.2010.00213.x
1
2
3
PE: Sharanya
Construction Morphology
CE: Shantha
4
5
6 Geert Booij*
7 University of Leiden
8
9
No. of pages: 13
10 Abstract
Toc head: Syntax & Morphology

11 In construction morphology, complex words are seen as constructions on the word level. The
12 notion ‘construction’, a pairing of form and meaning, as developed in the theory of Construction
13 Grammar, is essential for an insightful account of the properties of complex words. Morphological
14 patterns can be represented as constructional schemas that express generalizations about sets of exist-
Dispatch: 9.3.10
15 ing complex words and word forms, and provide the recipes for coining new (forms of) words. Such
16 schemas form part of a hierarchical lexicon with generalizations on different levels of abstraction,
17 they account for holistic properties of complex words that are not derivable from their constituents,
and they can be unified into complex schemas that express the co-occurrence of certain types of
18
word formation. The format of constructional schemas is also appropriate for phrasal lexical units
19 with word-like functions such as phrasal names, particle verbs, and periphrastic expressions.
20
21
22
B
23 1. Introduction
24 The theory of construction morphology (CM) aims at a better understanding of the
3
25 relation between syntax, morphology, and the lexicon, and of the semantic properties of
Manuscript No.
26
1
complex words. It provides a framework in which both the differences and the common-
27 alities of word level constructs and phrase level constructs can be accounted for.
2
28 There are two basic approaches to the linguistic analysis of complex words. In the
29 morpheme-based approach, which was dominant in post-Bloomfieldian American linguis-
30 tics, a complex word is seen as a concatenation of morphemes. In this approach, morpho-
31 logical analysis is conceived of as the ‘syntax of morphemes’. For instance, the English
32 word awareness can be analyzed as the concatenation of the adjectival morpheme aware
33
L N C 3
Journal Name
and the nominalizing suffix -ness that evokes the meaning ‘state, property’. Similarly, the
34 past tense form walked is analyzed as the concatenation of the morphemes walk and -ed.
35 This tradition of morphological analysis is manifest in the theory of Distributed Morpho-
36 logy (Harley and Noyer 1999).
37 Alternatively, we might take a word-based perspective in which words are the starting
38 points of morphological analysis. In this kind of morphological analysis, we compare sets
39 of words like the following:
40
41 (1) a. bald b. baldness
42 big bigness
43 black blackness
44 British Britishness
45 We then conclude to a formal difference between the words in (1a) and those in (1b)
46 that correlates systematically with a meaning difference: the words in (1b) have an addi-
47 tional sequence -ness compared to those in (1a) and denote the property or state
48 expressed by the adjectives (1a). This paradigmatic relationship between these sets of
49
ª 2010 The Author

Journal Compilation ª 2010 Blackwell Publishing Ltd
2 Geert Booij
1 words can be projected onto the word awareness in the form of word-internal morpho-
2 logical structure:
3
(2) [[aware]A ness]N
4
5 Moreover, the set of words in (1) may give rise to an abstract schema of the following
6 form in the mind of the speaker of English:
7
(3) [[x]A ness]N ‘the property ⁄ state of A’
8
9 This schema expresses a generalization about the form and meaning of existing deadjecti-
10 val nouns in -ness listed in the English lexicon, and also functions as the starting point for
11 coining new English nouns in -ness. That is, new deverbal nouns in -ness are not neces-
12 sarily coined on analogy with a specific existing word in -ness, but may be formed on the
13 basis of this abstract schema. A new word is formed by replacing the variable x in the
14 schema with a concrete adjective. For instance, the unification of the adjective [carless]A
15 with schema (3) results in the word construct [[carless]Vness]N with the meaning ‘the state
16 of being without a car’ (source: Time, October 5, 2009). That is, through unification the
17 variables in the formal structure and the semantic specification of the schema are turned
18 into constants. Unification is the basic operation, both at the word level and the phrase
19 level, to create well formed linguistic expressions.
20 The idea that word formation patterns can be seen as abstractions across sets of related
21 words is rooted in a venerable tradition. For instance, the German linguist Hermann Paul
22 wrote in his famous Prinzipien der Sprachgeschichte, published in 1880, that the language
23 learner will start with learning individual words and word forms, but will gradually
24 abstract away from the concrete words (s)he has learned, and coin new words and word
25 forms according to abstract schemas. This enables the language user to be creative both in
26 word formation and in inflection (Paul 1880 [3rd edition 1898]). This tradition is contin-
27 ued in the paradigmatic approach to word formation (Schultink 1962; Van Marle 1985),
28 and in recent work in varieties of non-transformational generative grammar such as
29 Head-driven Phrase Structure Grammar (HPSG) (Riehemann 1998, 2001).
30 As such morphological schemas depend on relationships between words, this morpho-
31 logical model has been called the network model (Bybee 1995), and the notion ‘network’
32 is indeed a proper term for conceptualizing the set of relationships between words in a
33 lexicon (Bochner 1993). Schema (3) may be said to license the individual nouns in -ness
34 in the English lexicon. Complex words, once coined, will be stored in the lexicon of a
35 language (which generalizes over the lexical memories of the individual speakers of that
36 language), if they have idiosyncratic properties and ⁄ or have become conventionalized.
37 CM assumes that complex words, i.e. the outputs of morphological operations, can be
38 listed in the lexicon. Morphological schemas therefore have two functions: they express
39 predictable properties of existing complex words and indicate how new ones can be
40 coined (Jackendoff 1975). This conception of the grammar avoids the well known rule
41 versus list fallacy (Langacker 1987), the unwarranted assumption that linguistic constructs
42 are either generated by rule or listed and that being listed excludes a linguistic construct
43 from being linked to a rule at the same time.
44 The relation between schema (3) and the individual words that conform to this schema
45 is that of ‘instantiation’: each of the nouns in -ness listed in (1) instantiate the schema in
46 (3). Schema (3) provides a direct account of the fact that -ness is a bound morpheme that
47 does not occur as a word by itself.
48 What is the implication of word-based morphology as outlined above for our concep-
49 tion of the architecture of the grammar? How does morphology fit into that architecture?
ª 2010 The Author Language and Linguistics Compass 3/1 (2010): 1–13, 10.1111/j.1749-818x.2010.00213.x
Construction Morphology 3 1
1 My starting point is that each word is a linguistic sign, a pairing of form, and meaning.
2 The form of a word in its turn comprises two dimensions, its phonological form, and its
3 morpho-syntactic properties. Hence, each word is a pairing of three types of information.
4 Morphology affects all three dimensions of words. That is why we need a ‘tripartite
5 parallel architecture’ of the grammar (Jackendoff 2002, 2007; Culicover and Jackendoff
6 2005, 2006). In sum, a word is a complex piece of information, and morphology deals
7 with the systematic pairing of form and meaning at the word level. In the next sections, I
8 will adduce a number of observations and arguments in favor of the claim that the notion
9 ‘construction’ and the related notion of ‘hierarchical lexicon’ are indispensable for an
10 insightful analysis of complex words. In this article I will focus on the relevance of the
11 constructional approach for word formation, but it is equally relevant for inflectional
12 phenomena, as I will briefly explain at the end of this article.
13
14
2. Constructions
15
16 The notion construction (defined as a pairing of form and meaning) is a traditional notion
17 used in thousands of linguistic articles and books. In most cases, it refers to a syntactic
18 pattern in which particular formal properties correlate with specific semantics that is not
19 completely compositional, but yet predictable. For instance, many linguists of English
20 speak of the passive construction because sentences with passive meaning in English have
21 a specific syntactic form that correlates with a specific passive meaning.
22 A famous example of a syntactic construction is the caused motion construction exem-
23 plified by sentence (4) (Goldberg 2006):
24
(4) Pat sneezed the foam off the cappucino
25
26 In this sentence, the verb to sneeze is used as a transitive verb, although it is normally an
27 intransitive verb. Its use as a transitive verb correlates with the presence of an object that
28 moves along a path specified by a PP. The transitivity of the verb to sneeze, and the
29 meaning component that the sneezing caused the foam to move is therefore to be seen as
30 a property of this construction as a whole.
31 The notion ‘construction’ plays an important role in a number of recent linguistic mod-
32 els: Construction Grammar (Goldberg 1995, 2006; Croft 2001; Fried and Östman 2004),
33 the Simpler Syntax Model (Culicover and Jackendoff 2005, 2006), Cognitive Linguistics
34 (Langacker 1999), and HPSG (Sag et al. 2003; Sag 2007). The following features of the
35 constructional approach are of high relevance for the further articulation of CM:
36
(5) Pieces of syntactic structure can be listed in the lexicon with associated meanings,
37
just as individual words are; these are the MEANINGFUL CONSTRUCTIONS of the
38
language.
39
Construction grammar makes no principled distinction between words and rules: a
40
lexical entry is more word-like to the extent that it is fully specified, and more
41
rule-like to the extent that it contains variables [...].
42
Lexical entries are arranged in an inheritance hierarchy. (Jackendoff 2008).
43
44 It should be clear by now that the notion ‘construction’ has relevance for the theory
45 of word structure because complex words, like syntactic constructs, are instantiations of
46 constructional schemas. The view that complex words instantiate morphological
47 constructions is also stated explicitly in Croft (2001), Goldberg (2006: 5), and Inkelas and
48 Zoll (2005). An example of a constructional analysis of prefixed words is the analysis
49 of English be-verbs in Petré and Cuyckens (2008). Yet, the investigation of the
4 Geert Booij
1 constructional aspects of word structure is still in its beginnings (Culicover and Jackendoff
2 2006). The next sections will present a number of arguments in favor of the construc-
3 tional approach to word formation, and its relevance for inflection will be briefly argued
4 for at the end of this article.
5
6
3. The hierarchical lexicon
7
8 Let us return to the schema for English deadjectival nouns in (3). This schema can be
9 qualified as a constructional idiom at the word level, that is, a word level construction
10 with one fixed position, that of the suffix. Constructional idioms are schemas in which
11 one or more positions are lexically fixed. For instance, in the English construction Ni after
12 Ni exemplified by year after year, book after book, etc. with the meaning ‘Ns in succession’
13 the preposition slot is lexically fixed as after whereas the N positions are variables. The
14 individual deadjectival nouns in -ness are morphological constructs that instantiate
15 construction schema (3). Each individual noun in -ness listed in the English lexicon is
16 dominated by this schema and inherits its predictable properties from schema (3) and
17 from its adjectival base word. Hence, if a listed complex word is completely regular, all
18 information concerning this word counts as redundant, except for the information that it
19 exists, that is, belongs to the lexical convention of English.
20 Schema (3) is a case of derivation, word formation by means of an affix. Patterns of
21 compounding, the other main type of word formation in English, can also be represented
22 straightforwardly as constructions, as illustrated in schema (6) for nominal compounds
23 which, like most English compounds, are right-headed:
24
(6) [[a]Xk [b]Ni]Nj M [SEMi with relation R to SEMk]j
25
26 This kind of notation is used in Jackendoff (2002). The double arrow symbolizes the
27 relationship between a particular form and a particular meaning. The variable X stands
28 for the major lexical categories (N, V, A, and P). The variables a and b in this schema
29 stand for arbitrary sound sequences. The variables i, j, and k stand for the lexical indexes
30 on the phonological, syntactic, and semantic (SEM) properties of words. The use of pho-
31 nological variables indicates that phonological information does not play a restrictive role
32 in this type of word formation. In (6), the meaning contribution of the compound
33 schema is specified, as morphology deals with the correlation between form and meaning
34 in sets of complex words. The nature of R, the semantic relation between the two parts
35 of a compound, is not specified in the schema but has to be determined for each individ-
36 ual compound on the basis of the meaning of the compound constituents, and encyclope-
37 dic and contextual knowledge (Downing 1977); for a discussion of the semantic
38 regularities involved, see Jackendoff (2009). Schema (6) thus specifies only the very gen-
39 eral meaning contribution of the compound construction that it establishes a semantic
40 relation of some sort between the two constituents, and also that the right constituent,
41 which is the formal head of a compound, is its semantic head as well: a towel rack, for
42 instance, is a kind of rack, not a kind of towel.
43 The following English compounds exemplify the various options defined by schema
44 (6):
45
(7) NN book shelf, desk top, towel rack
46
VN drawbridge, pull tab
47
AN blackbird, greenhouse
48
PN afterthought, overdose, inland
49
1 The four patterns listed in (7) are subcases of schema (6). They differ in certain ways. For
2 instance, whereas in NN compounding the modifier N can be a compound itself, this is
3 not the case for AN compounds:
4
(8) NN recursive modifier: [[reference book] [shelf]],[[kitchen towel] [rack]]
5
AN recursive modifier: [[snow white] [book]], [light-green] [house]
6
7 Therefore, we need subschemas of (6) in which such specific restrictions are specified.
8 A clear advantage of this representation of English nominal compounds is that we do
9 not need a Right-hand Head Rule (Williams 1981) to express the generalization that the
10 word class of an English nominal compound is the same as that of its right constituent.
11 Schema (6) will in its turn be dominated by a more general right-headed schema for all
12 English compounds including those with a verbal or adjectival head (brain-wash, light-
13 green), in which the head position is specified as Y (Y = N, V, A). Thus, the necessity of
14 both schemas and subschemas for English compounds illustrates the importance of the
15 notion ‘hierarchical lexicon’ for morphological analysis.
16
17
4. Holistic properties of morphological constructions
18
19 An important argument for using the notion ‘morphological construction’ is that it
20 enables us to specify predictable semantic properties of sets of derived words that cannot
21 be deduced from the semantic properties of their constituent parts. An example is the use
22 of full reduplication for the expression of the plural meaning on nouns in Malay:
23
(9) ana ‘child’ ana-ana ‘children’
24
rumah ‘house’ rumah-rumah ‘houses’
25
26 In such reduplication constructions, the notion ‘plurality’ is not expressed by one of the
27 constituents of the plural noun; it is the construction as such, a configuration with two
28 identical constituents, that evokes this meaning.
29 Another example comes from Romance languages. French, Italian, and Spanish have
30 nominal compounds of the form VN such as:
31
(10) a. French
32
chauffe-eau ‘water heater’
33
coupe-ongles ‘nail clipper’
34
garde-barrière ‘gate keeper’
35
grille-pain ‘toaster’
36
b. Italian
37
lava-piatti ‘dish washer’
38
mangia-patate ‘potato eater’
39
porta-lettere ‘postman’
40
rompi-capo ‘brain teaser, puzzle’
41
c. Spanish
42
lanza-cohetes ‘rocket launcher’
43
come-curas ‘lit. eat priests, anti-clerical’
44
mata-sanes ‘lit. kill healthy people, quack doctor’
45
limpia-botas ‘lit. clean boots, boot black’
46
47 These VN compounds are all nominal compounds, consisting of a verbal stem followed
48 by a noun in either the singular or the plural form. These are exocentric compounds, as
49 the noun on the right is neither the formal nor the semantic head of the compound. For
6 Geert Booij
1 instance, the Italian compound word lava-piatti does not denote a certain type of piatti
2 ‘plates’, but an instrument that washes dishes. So there is no constituent to which the
3 meaning component ‘agent ⁄ instrument’ of these compounds can be assigned, even
4 though this meaning component is systematically present in these compounds. This is
5 why one finds analyses in the linguistic literature in which a nominalizing zero-suffix is
6 postulated, on analogy with overt agentive ⁄ instrumental noun-creating suffixes such as
7 English deverbal -er. The problem of such analyses is that there is no other motivation
8 for such zero-elements than the agent ⁄ instrument meaning, and the fact that the relevant
9 complex words are nouns. The position of such a zero-affix (is it a prefix or a suffix?) is
10 completely arbitrary. In a constructional analysis, the agent ⁄ instrument meaning is speci-
11 fied as a semantic property of the VN construction as a whole. Thus, the following
12 schema can be assumed for such Romance VN compounds:
13
(11) [[Vk][Ni]Nj M [AGENT ⁄ INSTRUMENTj OF ACTIONk ON OBJECTI]j
14
15 Schema (11) represents a morphological construction in which a specific morphological
16 form (exocentric compounds of the form VN) correlates with a non-compositional, but
17 predictable meaning. An additional predictable but non-compositional property of the
18 French exocentric VN-compounds is that they have masculine gender, irrespective of the
19 gender of the N-constituent. Exocentric compounds thus provide a strong argument in
20 favor of a constructional analysis of word formation.
21
22
5. Semantic subpatterns
23
24 The morphological schemas introduced above form part of a hierarchical lexicon, in
25 which schemas dominate individual complex words. By default, complex words inherit
26 the information specified in a schema, but a particular piece of information may be over-
27 ruled by an individual lexical item that instantiates that schema. For instance, the Dutch
28 suffix -baar ‘-able’ attaches to transitive verbs to form adjectives with the meaning ‘can
29 be V-ed’, for instance lees-baar ‘read-able’ derived from the transitive verb lees ‘to read’.
30 Yet, there are a few adjectives in -baar attached to intransitive verbs, such as werk-baar
31 ‘work-able’ derived from the intransitive verb werk ‘to work’. By making use of the
32 notion of default inheritance (Briscoe et al. 1993; Kilbury et al. 2006), we allow for
33 exceptional properties of words to be expressed without giving up the generalizations
34 that hold for most words of that class. In the specification of werkbaar as an existing
35 adjective of Dutch, the inherited specification that its verbal base is a transitive verb is
36 overruled.
37 In the domain of compounding, we also need subschemas because certain words may
38 receive a specific interpretation when they form part of a compound that they do not
39 have when used as independent words. This is, for instance, the case for a number of
40 nouns in Dutch NA compounds that have an intensifier meaning:
41
(12) Intensifying lexemes in Dutch X A compounds
42
Noun Example
43
ber-e ‘bear’ bere-sterk ‘very strong’, bere-aardig ‘very kind’
44
bloed ‘blood’ bloed-serieus ‘very serious’, bloed-link ‘very risky’
45
dood ‘death’ dood-eng ‘very scary’, dood-gewoon ‘very ordinary’
46
kei ‘boulder’ kei-goed ‘very good’, kei-gaaf ‘very nice’
47
pis ‘piss’ pis-nijdig ‘very angry’, pis-woedend ‘very angry’
48
poep ‘shit’ poep-heet ‘very hot’, poep-lekker ‘very pleasant’
49
1 ret-e ‘ass’ rete-leuk ‘very nice’, rete-spannend ‘very exciting’

2 reuz-e ‘giant’ reuze-leuk ‘very nice’, reuze-tof ‘very good’
3
4 (The -e’s in ber-e, ret-e, and reuz-e denote schwas that function as a linking element.) This
5 productive use of lexemes as intensifiers can be expressed by subschemas for adjectival
6 compounding in which the first position is lexically fixed. Hence, these are constructional
7 idioms at the word level, such as:
8 (13) [[bere]N [x]Ai]A M’very SEMi’
9
10 Schema (13) is a subschema of the NA compounding schema for Dutch; it will inherit all
11 properties of this general schema such as its right-headedness. However, the normal mean-
12 ing contribution of the modifier noun ber-e ‘bear’ is overruled by the meaning of the modi-
13 fying constituent as specified in (13), in accordance with the principle of default inheritance.
14 This type of semantic development can be found in many languages. Here is an
15 example from Maale, a North Omotic language spoken in Southern Ethiopia. The noun
16 nayi ‘child’ has developed the general meaning ‘agent’, as illustrated by the following
17 complex words (Amha 2001):
18 (14) a. bayi nayi
19 cattle child
20 ‘one who brings cattle to the grazing area’
21 b. waari nayi
22 goat child
23 ‘one who takes care of goats’
24 c. móótsi naya
25 cattle.camp child
26 ‘one who lives in a cattle camp and takes care of cattle there’
27
28 Because cattle herding is historically a task of children in the Maale-speaking society, the
29 word for child has acquired a more general agent meaning. Therefore, a specific compound
30 schema with naya as its right constituent, and with this agent meaning is required.
31
32 6. Schema unification
33
34 The Dutch deverbal adjectives in -baar ‘-able’ mentioned above form a productive deriva-
35 tional category, which can be subsequently prefixed with the negative prefix on- ‘un-’. In
36 many cases, the intermediate adjective is only a possible word, and not listed in the lexi-
37 con. This is the case for, among many others, the following adjectives:
38 (15) Verb Deverbal adjective On-adjective
39 bedwing ‘suppress’ bedwing-baar ‘suppressable’ on-bedwing-baar
40 ‘unsuppressable’
41 bestel ‘deliver’ bestel-baar ‘deliverable’ on-bestel-baar
42 ‘undeliverable’
43 blus ‘extinguish’ blus-baar ‘extinguishable’ on-blus-baar
44 ‘unextinguishable’
45
46 This pattern suggests that two word formation schemas can be unified into one complex
47 schema that licenses multiply complex adjectives without the existence of the intermedi-
48 ate positive adjective being required. The data in (15) imply that the following schema
49 unification applies:
8 Geert Booij
1
2 (16) [on A]A + [Vbaar]A = [on[[Vbaar]A]A
3 The unification of word formation templates accounts for the possibility of simultaneous
4 use of two or more word formation patterns (in the example above the formation of
5 deverbal adjectives and on-adjectives). The availability of such unified templates is the
6 result of the language user’s ability to establish a direct relation between a base word and
7 a complex word that is two or more derivational steps away from that base word. Such
8 unified schemas do not complicate the grammar, because their properties follow from the
9 unification of independently established word formation schemas. Thus, language users
10 may coin a new multiply complex negative adjective such as Dutch onbedwingbaar ‘unsup-
11 pressable’ directly from a verbal base bedwing ‘suppress’ without an intermediate step.
12 An example of the use of a unified word formation schema from English is the simul-
13 taneous attachment of the prefix de(s) and the verbalizing suffix -ate or -ize to nouns or
14 adjectives, as in:
15
16 (17) Noun Verb
17 caffeine de-caffein-ate
18 moral de-moral-ize
19 mythology de-mytholog-ize
20 nuclear de-nuclear-ize
21 Stalin de-stalin-ize
22 An intermediate verb like to stalinize is certainly a possible verb. Yet, we should not
23 require the existence of this verb as a necessary intermediate step in the coining of desta-
24 linization, as it is not the case that the use of the verb destalinize presupposes that the
25 object involved has first been subject to a process of stalinization. That is, we assume a
26 unified template of the following form for verbs such as destalinize:
27
28 (18) [de [[x]Ni ize]V]V M REMOVE PROPERTY RELATED TO SEMi
29 In sum, by representing word formation processes as constructional schemas that can be
30 unified, it is possible to express that a multiply complex word can be derived in one step
31 from a base word that is two degrees less complex.
32
33
34 7. Word-like phrasal expressions
35 The lexicon is the repository of all simplex words and of all complex words that are idio-
36 syncratic and ⁄ or conventionalized. In addition, the lexicon has to specify multi-word
37 units that are idiomatic. The unpredictable properties of a linguistic construct have to be
38 learned and memorized by the speaker. The size of idiomatic constructs may vary from
39 sentences (for instance, proverbs) to phrases consisting of two words, the minimal size for
40 lexical phrases (for instance, the NP red tape as idiom for bureaucracy, or black death for
41 ‘pest’).
42 Lexical units may be construed productively by means of syntactic principles, although
43 they are word-like. Such constructs are sometimes referred to as ‘loose compounds’. The
44 advantage of a constructional approach to the analysis of such lexical units is that the
45 similarities with complex words can be expressed, without losing sight of the fact that
46 they reflect the syntactic principles of the language involved. Such loose compounds are
47 characteristic of Romance languages. For instance, the following French phrases are all
48 used as lexical units (Fradin 2003):
49
1 (19) a. N de N: fil de fer ‘iron wire’

2 b. N à N: moulin à vent ‘wind mill’
3 c. N à Det N: sauce à l’ail ‘garlic sauce’
4 d. AN : moyen âge ‘Middle Ages’
5 e. NA: poids lourds ‘heavyweight’
6
7 The patterns exemplified in (19) have a certain degree of productivity. In particular, the
8 French construction N à N is very productive for coining new names for objects, as illus-
9 trated in (20):
10 (20) moulin à poivre ‘pepper mill’
11 verre à vin ‘wine glas’
12 bois à feu’firewood’
13 fruit à confiture ‘jam fruit’
14 moteur à essence ‘petrol engine’
15
16 Note the difference between verre à vin ‘wine glass’ and verre de vin ‘glass of wine’. The
17 constructs with à have typically the role of classifying labels for entities.
18 As has been pointed out in the recent literature, multi-word expressions (MWEs) are
19 not just fixed sequences of words with an atomic meaning but differ in their degree of
20 compositionality and syntactic flexibility (Pitt and Katz 2000; Sag et al. 2002). The notion
21 ‘constructional idiom’ introduced above can be used to do justice to certain aspects of
22 this flexibility, in particular to the fact that idiomatic constructions can receive new
23 instantiations. The N á N construction reflects the syntax of French: in French NPs, the
24 head N precedes its complement, and PP complements begin with a preposition, such as
25 à. Yet, the N à N constructs are special in that the preposition is followed by a bare
26 noun, whereas normally in a PP, the N must be preceded by a determiner. The use of
27 bare nouns is tied to using nouns for non-referential, classificatory purposes. In sum, the
28 N à N construction is a subschema of the French NP construction [N PP], with the
29 specific properties that the preposition is fixed as à, and the complement of the preposi-
30 tion is a bare noun. The semantics of the construction is similar to that for the English
31 nominal compounds discussed above. The semantic role of the preposition à is establish-
32 ing some semantic relationship between the head noun on the left and the right noun
33 that functions as a modifier.
34 Another class of word-like phrasal units is the phrasal verbs of Germanic languages,
35 usually referred to as particle verbs. Examples from English are to put down and to phone
36 up. The Dutch equivalents of these particle verbs are neer-leggen and op-bellen, respec-
37 tively, with the particle preceding the verb (written as one word, even though they are
38 phrasal and other words can come in between the particle and the verb) (Blom and
39 Booij 2003; Blom 2005a,b). The separability of these particles is illustrated by the
40 following example:
41 (21) a. Ik hoorde dat Jan zijn moeder op belde
42 I heard that John his mother up phoned
43 ‘I heard that John phoned his mother’
44 b. Jan belde zijn moeder op
45 John phoned his mother up
46 ‘John phoned his mother’
47
48 In the embedded clause in (21a), the particle appears right in front of the verb; but in
49 main clause (21b), the finite verb has to appear in second position, whereas the particle
10 Geert Booij
1 occurs at the end of the sentence. This shows that particle verbs are phrasal in nature, in
2 accordance with the principle of Lexical Integrity that syntactic rules cannot move parts
3 of words. Hence, we have to assume phrasal constructional schemas for the various types
4 of particle verbs, in which the specific meaning contribution of each particle is specified.
5 For instance, the specific meaning of the word op ‘up’ used as a particle is that of cogni-
6 tive activation:
7
(22) [[op]P [x]Vi]V’ M ACTIVATE COGNITIVELY BY SEMi
8
9 (V’ indicates a syntactic projection of V, expressing that particle verbs are minimal
10 phrases.)
11 These particle verbs function as alternatives for prefixation in the coinage of complex
12 predicates, and this explains the restricted productivity of deverbal prefixation in
13 Germanic languages: there is strong competition from particle verb formation which is a
14 functionally equivalent means of creating complex predicates. Particle verbs can thus be
15 seen as instantiations of phrasal constructional idioms, whereas prefixed verbs are instanti-
16 ations of constructional idioms at the word level. This distinction is illustrated here by
17 means of the following minimal pairs from Dutch:
18
(23) Particle verb Prefixed verb
19
over komen ‘to come over’ over-komen ‘to happen to’
20
door leven ‘to continue living’ door-leven ‘to live through’
21
22 Prefixed verbs are not split in main clauses, unlike particle verbs. Hence, the difference
23 between the following two sentences:
24
(24) a. Jan komt het weekend over
25
John comes the weekend over
26
‘John comes over for the weekend’
27
b. Jan over-kwam een ongeluk
28
John over-came an accident
29
‘An accident happened to John’
30
31 In sum, particle verbs are lexical, yet phrasal units, and we can do justice to their proper-
32 ties by analyzing them as being formed according to phrasal constructional schemas. By
33 using the notion ‘constructional idiom’ for the analysis of particle verbs, we can maintain
34 the boundary between phrasal and morphological constructs, and yet do justice to the
35 word-like properties of particle verbs.
36
37
8. Inflection
38
39 Inflectional phenomena provide strong arguments for the constructional approach. The
40 classical problem of inflectional morphology is the complicated relation between form
41 and meaning. It is often impossible to assign a specific meaning to an inflectional affix,
42 because its actual value depends on the kind of stem it combines with, and the properties
43 of that stem, unless one allows for large sets of homonymous inflectional affixes. Con-
44 sider, for instance, the paradigm of masculine and neuter nouns (declension I) in Russian
45 (Gurevich 2006: 51):
46 (25) Masculine Neuter
47 SG PL SG PL
48 NOM stol stol-y bljud-o bljud-a
49
1 ACC stol-a stol-y bljud-a bljud-a

2 GEN stol-a stol-ov bljud-a bljud
3 DAT stol-u stol-am bljud-u bljud-am
4 INST stol-om stol-ami bljud-om bljud-ami
5 LOC stol-e stol-ax bljud-e bljud-ax
6 ‘table’ ‘dish’
7
8 As these paradigms illustrate, the same ending, for instance -a, may have different inter-
9 pretations depending on the class of the noun. Moreover, the particular value expressed
10 is a combination of properties, such as [ACC.SG] or [NOM.PL]. That is, there is no
11 one-to-one correspondence between form and morpho-syntactic properties. One also
12 finds elements in inflectional forms such as the thematic vowels of verbal conjugation
13 in Latin and the Romance languages that do not contribute by themselves to the mean-
14 ing of the inflected forms; they are ‘morphomic’ properties (Aronoff 1994). Hence, the
15 morpho-syntactic properties of each word form in the paradigm are best considered as
16 constructional properties, that is, as properties of the word form as a whole. This may
17 be expressed by morphological schemas that abstract, for instance, over words of
18 the same declension class such as the Russian ACC ⁄ GEN.SG word forms stola and bljuda
19 in (25):
20 (26) (x-a)xi M [N]i, masc.sg, acc ⁄ genMSEMi
21
22 where x is a phonological variable for nominal stems, and x is a phonological word. The
23 meaning SEMi mentioned here is that of the lexeme. The semantic interpretation of the
24 morpho-syntactic features is not specified here, because this interpretation depends on the
25 syntactic contexts in which a word occurs.
26 Another argument for the constructional approach to inflection is formed by periphras-
27 tic expressions. For instance, in English the word combination ‘have + past participle’ is
28 used to express the perfect tense of verbs. In this construction, the verb have does not
29 express the meaning ‘to possess’, but a grammatical meaning of perfectivity, in combina-
30 tion with the past participle. The grammatical meaning of perfect tense is a property of
31 this construction as a whole (Sadler and Spencer 2001; Ackerman and Stump 2004; Spen-
32 cer 2004). Hence, periphrastic inflectional forms are to be treated as constructional idioms
33 in which the auxiliary is lexically fixed, whereas the slot for the participle is a variable.
34 The semantic properties of periphrastic forms, such as perfectivity, are holistic properties
35 that are specified as properties of the construction.
36
37 9. Conclusion
38
39 In this article, I have argued that the notion ‘construction’ should be used for insightful
40 analyses of morphological phenomena. In this approach, the distinction between syntax
41 and morphology is maintained. Yet, the similarities between syntactic and morphological
42 constructs can be expressed as well. In particular, we find constructional idioms at both
43 the syntactic and the word level and thus, we can account for the word-like function of
44 productive phrasal constructions such as the ‘loose compounds’ of Romance languages
45 and the particle verbs of Germanic languages. Constructional schemas form part of a
46 hierarchical lexicon, which makes it possible to express subgeneralizations about sets of
47 complex words without obliterating the properties they share with other complex
48 words.
49
12 Geert Booij
1 Short Biography
2
Geert Booij (1947) is professor of Linguistics at the University of Leiden, where he holds
3
the chair for ‘Morphology and the lexicon’. His present research focuses on the structure
4
of the lexicon and the place of morphology in the architecture of the grammar. After his
5
BA and MA studies in Dutch and general linguistics at the University of Groningen (1965–
6
1971), he obtained his PhD degree at the University of Amsterdam (1977) where he taught
7
until 1981. From 1981–2005, he was professor of general linguistics at the Vrije Universi-
8
teit Amsterdam. He published a number of scholarly articles in linguistic journals and books
9
on phonology and morphology, and three books at Oxford University Press (The phonology
10
of Dutch, 1995; The morphology of Dutch, 2002; and the textbook The grammar of words, 2005,
11
20072). Together with Christian Lehmann and Joachim Mugdan, he edited the Handbook 2
12
on Inflection and Word formation (2 vols, 2000 ⁄ 2004, published by De Gruyter, Berlin). He is
13
the founder and editor of the Yearbook of Morphology (1988–2005) and its successor, the
14
journal Morphology. In 2010, his monograph Construction morphology will appear at Oxford
15
University Press. He served as the Dean of the Faculty of Letters of the Vrije Universiteit
16
Amsterdam (1988–1991, 1998–2002) and the University of Leiden (2005–2007), and as
17
member and chair of the Dutch Research Council for the Humanities (1997–2004).
18
Homepage with downloadable publications: http://www.hum2.leidenuniv.nl/booijge/. 3
19
20
21 Notes
22 * Correspondence address: Geert Booij, Faculty of Humanities, Leiden University Centre of Linguistics, University
23 of Leiden, PO Box 9515, 2300 RA Leiden, The Netherlands. E-mail: [email protected]
24
25
26
27 Works Cited
28 Ackerman, Farrell, and Gregory S. Stump. 2004. Paradigms and periphrastic expression: a study in realization-based
29 lexicalism. Projecting morphology, ed. by Louisa Sadler and Andrew Spencer, 111–57. Stanford: CSLI.
Amha, Azeb. 2001. The Maale language. Leiden: University of Leiden, Research School of Asian, African, and
30 Amerindian Studies.
31 Aronoff, Mark. 1994. Morphology by itself: stems and inflectional classes. Cambridge Mass: MIT Press.
32 Blom, Corrien. 2005a. The demarcation of morphology and syntax: a diachronic perspective on particle verbs.
33 Morphology and its demarcations, ed. by Wolfgang U. Dressler, Dieter Kastovsky, Oskar E. Pfeiffer and Franz
Rainer, 53–66. Amsterdam ⁄ Philadelphia: Benjamins. 4
34 ——. 2005b. Complex predicates in Dutch. Synchrony and diachrony. Utrecht: LOT.
35 ——, and Geert Booij. 2003. The diachrony of complex predicates in Dutch: a case study in grammaticalization.
36 Acta Linguistica Hungarica 50. 61–91.
Bochner, Harry. 1993. Simplicity in generative morphology. Berlin ⁄ New York: Mouton de Gruyter.
37 Briscoe, Edward, Ann Copestake, and Valeria de Paiva (eds). 1993. Inheritance, defaults and the lexicon.
38 Cambridge: Cambridge University Press.
39 Bybee, Joan. 1995. Regular morphology and the lexicon. Language and Cognitive Processes 10. 425–55.
40 Croft, William. 2001. Radical construction grammar. Syntactic theory in typological perspective. Oxford: Oxford
University Press.
41 Culicover, Peter W., and Ray Jackendoff. 2005. Simpler syntax. Oxford: Oxford University Press.
42 ——, and ——. 2006. The simpler syntax hypothesis. Trends in Cognitive Science 10. 413–8.
43 Downing, Pamela. 1977. On the creation and use of English compound nouns. Language 53. 810–42.
44 Fradin, Bernard. 2003. Nouvelles approches en morphologie. Paris: Presses Universitaires de France.
Fried, Mirjam, and Jan-Ola Östman. 2004. Construction grammar: a thumbnail sketch. Construction grammar in a
45 cross-linguistic perspective, ed. by Mirjam Fried and Jan-Ola Östman, ????–????. . Amsterdam ⁄ Philadelphia:
46 Benjamins. 5
47 Goldberg, Adele. 1995. Constructions. A construction grammar approach to argument structure. Chicago: Chicago
48 University Press.
——. 2006. Constructions at work. The nature of generalization in language. Oxford: Oxford University Press. .
49
1 Gurevich, Olga. 2006. Constructional morphology: the Georgian version, PhD dissertation. Stanford: Stanford
2 University.
Harley, Heidi, and Rolf Noyer. 1999. Distributed morphology. Glot International 4. 3–9.
3 Inkelas, Sharon, and Cheryll Zoll. 2005. Reduplication. Doubling in morphology. Cambridge: Cambridge Univer-
4 sity Press.
5 Jackendoff, Ray. 1975. Semantic and morphological regularities in the lexicon. Language 51. 639–71.
——. 2002. Foundations of language. Oxford: Oxford University Press.
6 ——. 2007. A parallel architecture perspective on language processing. Brain Research 1146. 2–22.
7 ——. 2008. Construction after construction and its theoretical challenge. Language 84. 8–28.
8 ——. 2009. Compounding in the parallel architecture and conceptual semantics. The Oxford handbook of
9 compounding, ed. by Rochelle Lieber and Pavol Stekauer, 105–29. Oxford, UK: Oxford University Press.
Kilbury, James, Wiebke Petersen, and Christoph Rumpf. 2006. Inheritance-based models of the lexicon. Advances
10 in the theory of the lexicon, ed. by Dieter Wunderlich, 429–80. Berlin: Mouton de Gruyter.
11 Langacker, Ronald. 1987. Foundations of cognitive grammar, vol. 1: theoretical prerequisites. Stanford, CA:
12 Stanford University Press.
13 ——. 1999. Grammar and conceptualization. Berlin: Mouton de Gruyter.
Paul, Hermann. 1880 [3rd edition 1898]. Prinzipien der Sprachgeschichte. Halle: Max Niemeyer.
14 Petré, Peter, and Hubert Cuyckens. 2008. Bedusted, yet not beheaded: the role of be-’s constructional properties in
15 its conservation. Constructions and language change, ed. by Alexander Bergs and Gabriele Diewald, 133–69.
16 Berlin ⁄ New York: Mouton de Gruyter.
17 Pitt, David, and Jerrold J. Katz. 2000. Compositional idioms. Language 76. 409–32.
Riehemann, Suzanne Z. 1998. Type-based derivational morphology. Journal of Comparative Germanic Linguistics
18 2. 49–77.
19 ——. 2001. A constructional approach to idioms and word formation, PhD dissertation. Stanford: Stanford Univer-
20 sity.
21 Sadler, Louisa, and Andrew Spencer. 2001. Syntax as an exponent of morphological features. Yearbook of
Morphology 2000, ed. by Geert Booij and Jaap Van Marle, 71–96. Dordrecht: Kluwer.
22 Sag, Ivan A. 2007. Sign-based construction grammar. An informal synopsis. Ms., Manuscript. Stanford, CA.
23 ——, Timothy Baldwin, Francis Bond, Ann Copestake, and Daniel Paul Flickinger. 2002. Multiword expressions:
24 a pain in the neck for NLP. Proceedings of CICLING 2002, ed. by Alexander Gelbukh, ????–????. Dordrecht:
Springer. 6
25 ——, Thomas Wasow, and Emily M. Bender. 2003. Syntactic theory. A formal introduction. Stanford, CA: CSLI
26 Publications.
27 Schultink, Henk. 1962. De morfologische valentie van het ongelede adjectief in modern Nederlands. Den Haag:
28 Van Goor Zonen.
Spencer, Andrew. 2004. Morphology – an overview of central concepts. Projecting morphology, ed. by Louisa
29 Sadler and Andrew Spencer, 67–109. Stanford: CSLI.
30 Van Marle, Jaap. 1985. On the paradigmatic dimension of morphological creativity. Dordrecht: Foris.
31 Williams, Edwin. 1981. On the notions ‘lexically related’ and ‘head of a word’. Linguistic Inquiry 12. 245–74.
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
Author Query Form
Journal: LNC3
Article: 213
Dear Author,
During the copy-editing of your paper, the following queries arose. Please respond to these by
marking up your proofs with the necessary changes/additions. Please write your answers on the
query sheet if there is insufficient space on the page proofs. Please write clearly and follow the
conventions shown on the attached corrections sheet. If returning the proof by fax do not write
too close to the paper’s edge. Please remember that illegible mark-ups may delay publication.
Many thanks for your assistance.
Query reference Query Remarks

Q1 AUTHOR: A running head short title was not
supplied; please check if this one is suitable
and, if not, please supply a short title of up to
40 characters that can be used instead.
Q2 AUTHOR: Please check the superscript num-
ber here.
Q3 AUTHOR: Please check this website address
and confirm that it is correct. (Please note that
it is the responsibility of the author(s) to
ensure that all URLs given in this article are
correct and useable).
Q4 AUTHOR: Please provide the actual place of
the publisher for references Blom (2005a),
Bochner (1993), Fried and Östman (2004), Pet-
ré and Cuyckens (2008).
Q5 AUTHOR: Please provide the page range for
reference Fried, and Östman (2004).
Q6 AUTHOR: Please provide the page range for
reference Sag et al. (2002).
USING E-ANNOTATION TOOLS FOR ELECTRONIC PROOF CORRECTION
Required Software
Adobe Acrobat Professional or Acrobat Reader (version 7.0 or above) is required to e-annotate PDFs.
Acrobat 8 Reader is a free download: http://www.adobe.com/products/acrobat/readstep2.html
Once you have Acrobat Reader 8 on your PC and open the proof, you will see the Commenting Toolbar (if it
does not appear automatically go to Tools>Commenting>Commenting Toolbar). The Commenting Toolbar
looks like this:
If you experience problems annotating files in Adobe Acrobat Reader 9 then you may need to change a
preference setting in order to edit.
In the “Documents” category under “Edit – Preferences”, please select the category ‘Documents’ and
change the setting “PDF/A mode:” to “Never”.
Note Tool — For making notes at specific points in the text

Marks a point on the paper where a note or question needs to be addressed.
How to use it:

1. Right click into area of either inserted
text or relevance to note
2. Select Add Note and a yellow speech
bubble symbol and text box will appear
3. Type comment into the text box
4. Click the X in the top right hand corner
of the note box to close.
Replacement text tool — For deleting one word/section of text and replacing it
Strikes red line through text and opens up a replacement text box.
How to use it:

1. Select cursor from toolbar
2. Highlight word or sentence
3. Right click
4. Select Replace Text (Comment) option
5. Type replacement text in blue box
6. Click outside of the blue box to close
Cross out text tool — For deleting text when there is nothing to replace selection
Strikes through text in a red line.
How to use it:
1. Select cursor from toolbar
2. Highlight word or sentence
3. Right click
4. Select Cross Out Text
Page 1 of 3
Approved tool — For approving a proof and that no corrections at all are required.
How to use it:

1. Click on the Stamp Tool in the toolbar
2. Select the Approved rubber stamp from
the ‘standard business’ selection
3. Click on the text where you want to rubber
stamp to appear (usually first page)
Highlight tool — For highlighting selection that should be changed to bold or italic.
Highlights text in yellow and opens up a text box.
How to use it:

1. Select Highlighter Tool from the
commenting toolbar
2. Highlight the desired text
3. Add a note detailing the required change
Attach File Tool — For inserting large amounts of text or replacement figures as a files.
Inserts symbol and speech bubble where a file has been inserted.
How to use it:

1. Click on paperclip icon in the commenting toolbar
2. Click where you want to insert the attachment
3. Select the saved file from your PC/network
4. Select appearance of icon (paperclip, graph, attachment or
tag) and close
Pencil tool — For circling parts of figures or making freeform marks

Creates freeform shapes with a pencil tool. Particularly with graphics within the proof it may be useful to use
the Drawing Markups toolbar. These tools allow you to draw circles, lines and comment on these marks.
How to use it:

1. Select Tools > Drawing Markups > Pencil Tool
2. Draw with the cursor
3. Multiple pieces of pencil annotation can be grouped together
4. Once finished, move the cursor over the shape until an arrowhead appears
and right click
5. Select Open Pop-Up Note and type in a details of required change
6. Click the X in the top right hand corner of the note box to close.
Page 2 of 3
View publication stats

Construction Morphology

Uploaded by

Copyright:

Available Formats

Construction Morphology

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Construction Morphology

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Article in Language and Linguistics Compass · July 2010

The user has requested enhancement of the downloaded file.

Toc head: Syntax & Morphology

ª 2010 The Author

1 ret-e ‘ass’ rete-leuk ‘very nice’, rete-spannend ‘very exciting’

1 (19) a. N de N: fil de fer ‘iron wire’

1 ACC stol-a stol-y bljud-a bljud-a

Many thanks for your assistance.

Query reference Query Remarks

Note Tool — For making notes at specific points in the text

How to use it:

How to use it:

How to use it:

How to use it:

How to use it:

Pencil tool — For circling parts of figures or making freeform marks

How to use it:

You might also like