Academic journal article SKASE Journal of Theoretical Linguistics

Blends at the Intersection of Addition and Subtraction: Evidence from Processing

Academic journal article SKASE Journal of Theoretical Linguistics

Blends at the Intersection of Addition and Subtraction: Evidence from Processing

Article excerpt

1. Introduction

Would you easily guess that a negatude is a negative attitude? It is likely that to unpack the blend negatude would be less difficult than to decompose the clipping compound finlit into its constituents financial and literacy. This difference can be explained by the fact that clipping compounds combine fragments of words which are too short to ensure the recognisability of their full counterparts (Gries 2006, 2012).

The representation of morphologically complex words has been conducted on material of various morphological categories. When we come across a compound word, even one we have never seen before, we can relate its meaning to the meanings of its constituents (e.g. predict that the meaning of juice bar has to do with juice and bar). When we come across a word which was formed with some degree of shortening of the constituents, the same task may become a lot more difficult. In fact, various situations are possible. On the one hand, two words, e.g. blizzard and disaster, can be blended together, as in blizzaster, so that some of the phonological and graphical material is lost and only SPLINTERS (in this case blizza- and -aster) are retained in the blend, or sometimes words can be blended by overlapping, as in predictionary [left arrow] prediction + dictionary. On the other hand, there is a case of clipping compounds, such as rumint [left arrow] rumour + intelligence, in which the clipped versions of the constituents, i.e. their beginnings enter the new formation. In some publications on word formation (e.g. in Adams (1973), Berg (1998) and others) no explicit distinction is made between blends and clipping compounds. Many researchers (e.g. Cannon (1986), Bertinetto (2001), Bauer (2006, 2012), to name just a few) classify them as two different word formation types. Most often this distinction is based on formal criteria, but sometimes, as in Kubozono (1990), semantic relations between the constituents (henceforth source words) of blends are taken into consideration.

This paper is in agreement with the idea suggested in Gries (2006) that blends and clipping compounds are formed according to different principles and differ in terms of similarity between the source words and their recognisability. This, in turn, suggests that novel words of these two categories should be processed differently. In Beliaeva (2014) a classification of such neologisms into several structural types was substantiated, and a differentiation between blends and clipping compounds was drawn on the basis of their semantics and origin. However, the differences in the recognisability of the constituents in blends and clipping compounds have not been tested experimentally. This paper reports on an experimental study in which blends and clipping compounds are compared in terms of the recognition and processing of their source words. Formal transparency (i.e. what portion and which part of the source words is preserved in the shortening) is studied as a factor determining the retrieval of the source words by the readers who are exposed to corresponding blends or clipping compounds.

Extensive experimental evidence for the storage and retrieval of words has been derived using the priming technique. In priming experiments, the response of participants to a stimulus referred to as the target is studied in relation to another stimulus presented before the target, the prime. The relatedness of prime to target (that is, whether the prime is identical to the target, phonologically or graphically, morphologically or semantically related to it, or unrelated) is manipulated in order to detect whether the primes which are related to targets in a particular way enhance or inhibit the participants' reaction to targets (Neely 1991; MarslenWilson et al. 1994). The aim of the present experiment is to investigate whether blends with a higher degree of formal transparency (such as predictionary above) may produce stronger priming of their source words in a lexical decision task than either blends with a lower degree of formal transparency such as blizzaster, or clipping compounds such as rumint. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.