Pernilla Hallonsten Halling

attr
Function: property modification within reference
Form: art ADJ N
Example: (4.7a)

preD

Function: property predication
Form: S be_cop ADJ Example: (4.7b)

aDv

Function: property modification within predication
Form 1: S V ADJ-ly _advz Form 2: S V ADJ_subset Example: (4.7c), (4.4c)
The Form lines in (4.16) contain the forms of whole construction, and not just individual roots, word forms, or lexemes. For attr, the construction is entirely schematic, composed of an article slot (art), an Adjective slot (ADJ) and a Noun slot (N). The construction in pred is partly substantive, since the Copula be is required. It is also schematic in the slots of the subject (S) and the Adjective (ADJ). In adv, (at least) two constructions must be captured as discussed above: one with Adverbs derived with -ly, and one with Adjectives used adverbially. But including many forms is not a problem for the constructional- typological approach. The labels in subscript in the constructional-typological notation provide a gloss for substantive items. In the case of be in (4.16b), this is glossed as a Copula, which can be compared to the Verb he in Lakota (see 4.17 below), glossed V to indicate that it is a lexical verb. Major elements are abbreviated in large capitals and include nouns, verbs, subjects, any type of modifiers, and stative verbs. Stative verbs are here defined as verbs that denote states, and not actions, and are often attested in languages that do not have adjectives. Conventions for notation other than glossing (which follows the glossing abbreviations listed in the beginning of this book) that are

used in the constructional-typological notation are listed in appendix A. When elements are schematic, they are abbreviated independently (e.g. art for ‘article’). When they are substantial, they are attached to the italicized schematic element in subscript (e.g. be_cop in pred in 4.16 above). Sometimes, translations are also provided for clarification. Full constructional notation for all languages can be found in appendix B.

An issue that arises at this point is how to delimit constructions. Starting with adv, in a fashion parallel to attr, it seems plausible to limit the construction found here to the modifier and the modified. For attr in English, this is an Adjective and a Noun that heads the NP, whereas for adv in English, this is an Adverb (whether derived or not) and a Verb. To capture the whole predication in which the modification is found, I have nonetheless chosen to include the Subject when it is overt, as in the case of English. Here, we may recall an important point from Role and Reference Grammar (see section 2.3.1), where adverbs denoting manner and pace are located in the periphery of the core. The core contains both the predicate and its arguments, and manner and pace adverbs are assumed to modify both. Only aspectual adverbs are located in the periphery of the nucleus, since they are assumed to modify the predicate alone. Also the semantic accounts discussed in chapter 2 pointed to various uses of manner adverbs and the subtle meaning distinctions that follow from these, interacting with the participants of an event. This provides further support for including any subject in the constructional notation for adv. For pred, the subject is included for the whole construction to be presented, on the same grounds as for adv. Alternatively, the English construction in pred may be denoted in the form of [NP be ADJ ] as proposed by Croft & Cruse (2004: 253). Although that is a reasonable option, I have nonetheless chosen to follow Koch (2012) more generally and use S for the subject slot. Another way in which delimitation of constructions may be problematic is when examples vary in their extension. For instance, in an English example such as Sheila drives her lawnmower fast, the object her lawnmower could also be included in its constructional-typological notation. But objects are usually not expected to affect the modification performed by property words in predicating expressions (as is evident in examples such as Sheila drives fast vs. Sheila drives her lawnmower fast ). Still, in the case of language-specific examples where the level of detail of the construction appears to be problematic in the typological study, I will discuss this. Note also that I do not assume any specific phrase structure in the constructional notation. It may seem counterintuitive if a different perspective is taken, where the V in the form lines of example (4.16) may be interpreted as the VP head. However, this is not implied by V in its present use. Rather, constructions are different from phrases. In terms of distinguishing between phrase-like components, I would rather argue for something along the lines of the idea of formal groupings proposed by Croft (2001: 190–197).
The notation in (4.16) shows how comparison is facilitated by the constructional- typological approach, even within a single language. The focus on the construction does not exclude the option of focusing on the lexeme, root, or form. The examples in (4.16) also illustrate that in a certain case (i.e. fast ), the same Adjective is used in all three functions, indicating an [attr pred adv] lexeme overlap for English. The constructional notation clearly spells out the similarities and differences both between forms within the same function (e.g. Form 1 and Form 2 in adv) and between forms within different functions (e.g. the one in pred and the two forms in adv). Recalling the examples from

Lakota in (4.8), the utility of the constructional notation can be made even clearer. If the encoding found for pred and adv is to be presented, the intermediate construction must be included in both functions, in addition to other constructions not discussed here (but see 7.18 in section 7.3.2). Note that only pred and adv are reproduced in (4.17), since these are the only examples from Lakota that have been described so far.

(4.17)
pred and adv in Lakota

preD

Function: property predication Form: S ST.V-ya/-yela_advz he_v Example: (4.8)

aDv

Function: property modification within predication
Form: S ST.V-ya/-yela_advz he_v
Example: (4.8)

Spelled out in this way, it becomes even clearer that Lakota has a construction that is between pred and adv, and must be represented in both functions (although it should again be noted that there are other constructions as well, more prototypical to the two functions). The Verb stems with the suffixes -ya and -yela are denoted as stative verbs (ST.V) in the constructional-typological notation.

So far, the benefits of the constructional-typological approach have been illustrated in terms of comparisons within the same language. But there are yet greater advantages when turning to cross-linguistic comparison, since it is possible to consistently capture any number of forms in one function and compare them to the forms in the same func- tion in another language (cf. Koch’s comparison of Brazilian Portuguese and French as reproduced in 4.15 above).
In summary, the constructional approach is highly suitable for typological comparison, since it is both comprehensive and allows for analysis of more limited domains. The nota- tion is a way of describing examples from individual languages in a much more useful way than giving a full description of each individual example. It provides a level of abstraction that allows for comparison, while still capturing the diversity of language-specific facts. It is not the only option for a typological study of this kind – other methods could also have been used. But it is difficult to picture any method that would perform as consis- tently and comprehensively, without designing a comparative standard from scratch. This strongly motivates the choice of the constructional-typological approach for the present study. While Construction Grammar is a family of theoretical models, it is not neces- sary to adhere to its theoretical claims in order to employ the constructional-typological method for descriptive purposes. Importantly, the latter is not intended to be a syntac- tic theory. However, constructions must naturally be acknowledged – which is generally done by typologists without much discussion – as well as the identified functions and the relevance of comparing them.

As this is not a theoretical syntax-oriented study, it is not about Construction Gram- mar. The constructional-typological method still serves as an appropriate tool for com- parison. Constructions are language-specific, and the constructional-typological method is employed for the purpose of comparing them. Therefore, it is within comparison that its strength lies.

4.