Pernilla Hallonsten Halling

y
there
a
have.prs.3sg
un
indf.m.sg
livre
book
sur
upon
la
def.f.sg
table
table

‘There is a book on the table.’
While the verb form tem in (4.13) is substantive, the remaining parts of the construction are schematic. Similarly, the complex verbal item y a in (4.14) is substantive, whereas the rest of the construction is schematic. Based on CxG notation, Koch denotes these two examples as in (4.15), with the direct object (DO) encoding the located (L^ED) and the locational adverbial (LOCA for short) encoding the locus (L^US) (2012: 551).
(4.15) Rhematic locational construction (Koch 2012: 551)

1. 
   Meaning: L^US locates L^ED Form: Braz. Portuguese LOCA ter _3sg DO Example: (4.13)
   

⁵ Koch (2012: 540–541) distinguishes rhematic location, e.g. There is a book on the table, from

thematic location, e.g. The book is on the table.

2. 
   Meaning: L^US locates L^ED
   

Form: French S^◦ LOCA y +avoir _3sg DO
Example: (4.14)
The notation in (4.15) captures the similarities between rhematic location in Brazilian Portuguese and French, namely the schematic items LOCA and DO, which are made up of noun phrases. The differences in substantive items also becomes clear, in that Brazilian Portuguese has the simple verb ter, whereas French has the complex verbal item y +avoir. Also, French syntax requires the dummy subject (S^◦) il. The French verbal item is captured without difficulty, showing that CxG is able to handle complex combinations (even idioms) in this notation. Both the verbal item, which is substantive, and the schematic noun phrases that go with it are taken into account, meaning that valency is not disregarded. More generally, the notation presents the constructions in their entirety, providing a good foundation for comparison within as well as across languages, without the risk of excluding any element. Koch elaborates the constructional-typological approach in a much more sophisticated way than it is employed here, to include other CxG feautures such as inheritance links (cf. 2012: 561ff.). But the description provided here should suffice to illustrate the utility of the constructional-typological approach for the present study.
Already in (4.9) in the previous section, I introduced the notation of the constructional- typological approach in the discussion of attr, pred, and adv. For the purpose of the present study, I have chosen to replace the term meaning with function. Constructions are usually defined as having meaning and form. In the case of attr, any attempt to capture its meaning would likely end up in something like what Hilpert (2014: 52) phrases as “an X has the quality of being Y”. This meaning may seem adequate at first, but will fail to cover many examples, such as old friends, which does not mean ‘friends that have the quality of being old’ (Hilpert 2014: 52), but rather ‘friends that have been friends for a long time’. In fact, as pointed out by Hilpert (2014: 52), Fillmore et al. (2012: 326) define the English construction in attr as a meaningless construction,⁶ and label it the Modifier-Head construction.⁷ According to Hilpert, the “construction reflects a formal generalisation, but it does not contribute any meaning of its own that would go beyond the combined
meanings of the component lexical items” (2014: 52). pred is no easier to pinpoint in
terms of meaning: ‘X is property Y’ or ‘X has the property of being Y’ are futile attempts. Along the same lines, proposing a meaning for adv such as ‘X performs action in Y way’ is too general. On the other hand, adv could perhaps be defined as a Modifier-Predicate construction, in a fashion parallel to the Modifier-Head construction in attr. Meaningless constructions in the account by Fillmore et al. (2012) are reminiscent of phrase structure rules, in that they only yield a grammatically acceptable combination of elements and nothing else. On the other hand, meaninglessness as a feature of these constructions seems uncalled for, especially against a CxG background, since this is an approach in which it is fundamental that constructions have meanings. Another perspective on constructions such as those found in attr, pred, and adv is that they have very general meanings, or even a family of meanings (cf. family resemblance according to Wittgenstein 1953). It
⁶ There are three meaningless constructions according to Fillmore et al. (2012: 326), the other two being
Head-Complement and Subject-Predicate.
⁷ This actually also covers adverbs modifying adjective, e.g. completely full (Fillmore et al. 2012: 326).

could also be argued that these constructions are extremely schematic, in the sense that very large classes of elements fit into their respective slots. Based on such considerations, I use the term function instead of meaning, to accurately describe these constructions. The definition of each function implements the discourse function for attr, pred, and adv following Croft (2001, see sections 2.3.4 and 3.2). The broad meanings of attr, pred, and adv are as such important to capture. Since they are constructions with very general meanings, it is more fitting to describe them as having functions than meanings. The term function will thus be used instead of meaning in the construction descriptions of attr, pred, and adv, even though this is not the way it would be done traditionally within CxG.

Let us reconsider the examples of English in constructional-typological notation from (4.9) in the previous section, reproduced in (4.16).
(4.16) attr, pred, and adv in English