Prepositions in UI( Monolingual Learners' Dictionaries ...

f.iu.-;uis1i,·s 2 3/ J I ).'J.-f ,-JO

MARIJA M. BRALA

(0 Oxford University Press 2002

FORUM

Prepositions in UI( Monolingual Learners' Dictionaries: Expanding on ·ndstro1nberg's Problems and Solutions MARIJA M. BRALA Lniversity of Trieste, Italy

In an insightful examination of the quality of lexicographic portrayals of prepositional semantics, Lindstromberg (2001) convincingly argues that prepositional representation in dictionaries is quite inadequate. Focusing his analysis on the preposition 'on', Lindstromberg shows that 'on' sense information in UK monolingual dictionaries is frequently inaccurate (vague, misleading, with mismatched examples), and concludes by suggesting some ways for improving entries on 'on', but also on spatial prepositions in general. While agrneing in principle with Lindstromberg's analysis, my view is that 1 1is Jrgumcnts arc not exhaustive, that is, in his paper the author fails to include some crucial psycholinguistic studies and relate his views to some current linguistic theories, which would undoubtedly make his thesis much more sound and convincing, possibly also providing a solid basis for better l('IH lt,cling recommendations. TilL.' first study that I see as being crucially related to Lindstromberg's work is a crosslinguistic study of prepositional usage by Bowerman and Pederson ( l 992; d. also Bowerman and Choi 2001: 484-7). In this detailed study the authors examine the physical (spatial) senses lexicalized by the English prepositions 'cm' and 'in', and the ways in which these same senses (i.e. types of spatial relations) are rendered in 33 other natural languages. Bowerman and Pederson aptly show that all the instances of spatial relations under consideration can be divided into 11 categories, with categorial boundaries being drawn whenever at least one language, in order to lexicalize one or rnorc of these spatial relations, 'switches' from one preposition (or other lexical form) 1 to another. Even more interestingly, the authors observe that rhcse categories can be ordered as to form the sequence shown in Figure 1. This ordered sequence of meaning categories, is, at a crosslinguistic level, differently partitioned into meaning clusters. For example, Spanish and 0on,1guese lexicalize rhe whole range with one preposition only ('en' and 'cm' respectively), English, uses two prepositions ('on' and 'in'), while Gnman and Dutch partition the scale into three 'prepositional segments' 1'aul', 'an', and 'in' for German, 'op', 'aan', and 'in' for Dutch), etc. The most

Support jMarks From 1 on a 1 surface below 1

I

i

E.g. Cup Writing on table ion paper

Clingy attach· ment

I Hanging I Fixed over/ 1 agamst

I attach-

I

I

mcnt

Rmndrops : Picture on Handle on Ion wall I j cupboard window

point attachment

Apple on

tivig

!i Encircle i Impaled! I Pierces with / contact I on Ring on

i Partiai

Apple an 1 Arrmv inf ()qarc;tc

I Stick

throu,qh

in mowh

135

'.Inclusion

ApFI(

Figure 1 The ON-IN scale of spatial meaning categories (Bowerman and Pederson 1992; Cf also Bowerman and Choi 2001) striking observation is that the portions of the scale attributed to different prepositions are 'compact', that is there is no language which would lexicalize part of the scale with 'on', then part of the scale by 'in', and then par! of the: scale by 'on' again. If there is overlapping at all (i.e. if a language uses two prepositions interchangeably for one or more categories) this always occurs in the section of the scale which is 'transitional', that is between the categories in which the use of only one of the two prepositions is possible.2 All this leads to the hypothesis that the ON-IN scale is not formed on a random basis, but that there must be an underlying 'gradient', something more powerful than 'linguistic arbitrariness' governing the formation and arrangement of its categories. Departing from this observation, in my Ph.D. research (Brala 2000), 1 se1 out to try to understand what this 'something' might be. After a literature review, extensive crosslinguistic probing, and two experimental case studies, my conclusion is that the categories of spatial relations are formed (and later organized into meaning clusters) on a combinatorial basis, out of universal, primitive, bodily-based semantic features. The latter are viewed as being shared between the human language faculty and other sub-systems of human cognition (cf. Talmy' 2000). What exactly is meant by this and how • can this be related to Lindstromberg's paper? Let us start by considering the categories on the 'on'-'in' scale not as topological (as, e,g., Bowerman and Lindstromberg do), but rather functional configurations (cf. Vandeloise 1998; also 1991). From this perspective, we get a new reading of prepositional semantics, and for the word class as such, we phrase the meaning of the relational lexical unit PREPOSITION as follows: in terms of which features does the Landmark control the location of the Located Object? The answer to this question determines the choice of the preposition. This simple formula easily explains the suggested perceptual difference in the construal of a Landmark noted by Lindstromberg (2001: 80) between 'frog in the grass', vs 'frog on the grass'; for 'in' to be a possible lexical choice, the Landmark needs to control the location of the LO in terms of voluminosity (see below), whereas for the English 'on' the Landmark controls the location of the LO in one of its (Landmark's) axes (usually the horizontal or the vertical). We thus have the perceptual 'adjustment' (or a specific conceptualization) of the