Volume 41, Issue 1 - Special issue: Semantic maps

In this paper, we present an overview of the methods associated with semantic maps, focusing on current challenges and new avenues for research in this area, which are at the core of the contributions to this special issue. Among the fundamental questions are: (1) the validity of the basic assumption, namely, to what extent does coexpression reflect semantic similarity; (2) the central problem of identifying analytical primitives in the domain of semantics; (3) the methods of inference used for creating coexpression maps and the representation techniques (graph structure vs. Euclidean space) as well as their respective merits (including the goodness of fit of the models); and (4) the use of semantic maps to support diachronic and synchronic descriptions of individual languages. In order to illustrate and discuss key aspects, we conduct an experiment in the semantic field of emotions, for which we construct a classical semantic map based on the dataset of CLICS 3 .

Although semantic maps and typological hierarchies are different analytical tools and make different predictions, there is, arguably, a particular kind of a semantic map that shares certain features with typological hierarchies, in particular, the property of directionality. First, we briefly illustrate that such maps are based on the notion of local markedness and are designed to model the interaction between grammatical categories. We then explore the Actionality Hierarchy, formulated in terms of Vendlerian classes, which models the interaction between actionality and grammatical aspect. On the one hand, it will be shown that the Actionality Hierarchy can be reconstructed as a semantic map, based on common semantic features shared selectively between individual Vendlerian classes. Notably, it is directional and can be used to capture relations of local markedness between actionality and aspect. On the other hand, we will provide first systematic, quantitative evidence for the Actionality Map, using cross-linguistic parallel corpus data of four languages.

We describe two mathematical representations for what have come to be called “semantic maps”, that is, representations of typological universals of linguistic co-expression with the aim of inferring similarity relations between concepts from those universals. The two mathematical representations are a graph structure and Euclidean space, the latter as inferred through multidimensional scaling. Graph structure representations come in two types. In both types, meanings are represented as vertices (nodes) and relations between meanings as edges (links). One representation is a pairwise co-expression graph, which represents all pairwise co-expression relations as edges in the graph; an example is CLICS. The other is a minimally connected co-expression graph – the “classic semantic map”. This represents only the edges necessary to maintain connectivity, that is, the principle that all the meanings expressed by a single form make up a connected subgraph of the whole graph. The Euclidean space represents meanings as points, and relations as Euclidean distance between points, in a specified number of spatial dimensions. We focus on the proper interpretation of both types of representations, algorithms for constructing the representations, measuring the goodness of fit of the representations to the data, and balancing goodness of fit with informativeness of the representation.

Whether it is based on philological data or on reconstruction, historical linguistics formulates etymological hypotheses that entail changes both in form and in meaning. Semantic change can be understood as a change in “patterns of lexification”, i. e., correspondences between forms and senses. Thus a polysemous word, which once lexified senses s1–s2–s3, evolves so it later encodes s2–s3–s4. Meanings that used to be colexified are now dislexified, and vice versa. Leaning on empirical data from Romance and from Oceanic, this study outlines a general model of historical lexicology, and identifies five types of structural innovations in the lexicon: split, merger, competition, shift, and relexification. The theoretical discussion is made easier by using a visual approach to structural change, in the form of diachronic maps. Semantic maps have already proven useful to represent synchronic patterns of lexification, outlining each language’s emic categories against a grid of etic senses. The same principle can be profitably used when analysing lexification patterns in diachrony: lexical change is then viewed as the reconfiguration of sense clusters in a semantic space. Maps help us visualize the “lexical tectonics” at play as words evolve over time, gradually shifting their meaning, gaining or losing semantic territory, colliding with each other, or disappearing forever.

Traditionally, lexical typology has to a large extent been interested in lexical categorization of various cognitive domains (e. g., colour , perception , body) , i. e., in how these are cut up by the most important words in a language, and in lexical motivation , or formal relatedness, i. e., in whether words for certain concepts are completely unrelated or related to others via polysemy or derivation (e. g., intransitive vs. transitive verbs, words for ‘day’ and ‘sun’, etc.). Grammatical behavior of words and morphosyntactic patterns as encoding meanings traditionally belong to grammatical typology. In this paper, I consider the domain of temperature and show how the close interaction between lexicon and grammar in the encoding of the temperature domain across languages calls for an integrated lexico-grammatical approach to these phenomena. As a useful tool for such an enterprise I suggest an elaborated semantic map comprising three layers – the layer specifying the words with the information on their mutual formal relations (i. e., whether they are identical, completely unrelated or related via derivation or inflection), their morphosyntactic properties (e. g., their part-of-speech affiliation, inflectional potential, etc.), and the constructions they occur in.

The present paper discusses connectivity and proximity maps of causative constructions and combines them with different types of typological data. In the first case study, I show how one can create a connectivity map based on a parallel corpus. This allows us to solve many problems, such as incomplete descriptions, inconsistent terminology and the problem of determining the semantic nodes. The second part focuses on proximity maps based on Multidimensional Scaling and compares the most important semantic distinctions, which are inferred from a parallel corpus of film subtitles and from grammar descriptions. The results suggest that corpus-based maps of tokens are more sensitive to cultural and genre-related differences in the prominence of specific causation scenarios than maps based on constructional types, which are described in reference grammars. The grammar-based maps also reveal a less clear structure, which can be due to incomplete semantic descriptions in grammars. Therefore, each approach has its shortcomings, which researchers need to be aware of.

This study is an exercise in micro-scale comparison based on the semantic map approach. A semantic map is constructed for meanings associated with the word for ‘grain’ in four closely related South Mande languages: Dan Blowo, Tura, Mwan, and Wan. Although non-cognate, the words show remarkable similarity in their range of meanings, suggesting that newly introduced words are associated with the same networks of meanings as the words they come to replace. Differences between the languages boil down to the unavailability or reduced productivity of particular uses in one or two of the languages (Wan and to some extent Mwan). The original network is shown to have become significantly reduced in Wan, and to a lesser extent in Mwan. A major challenge in micro-comparison is due to the small sample size and the idiosyncratic nature of some of the semantic relationships, which together make it hard to explore the network’s internal structure. Language-internal evidence can occasionally be used to compensate for the missing evidence from large-scale typological sampling. Overall, the approach proves promising for research on the history of closely related languages and has the potential to help lexicographers and fieldworkers identify possible gaps in lexical descriptions.

The paper outlines the basics of data collection, analysis and visualization under the frame-based approach to lexical typology and illustrates its methodology using the data of cross-linguistic research on verbs of falling. The framework reveals several challenges to semantic map modelling that usually escape researchers’ attention. These are: (1) principles of establishing lexical comparative concepts; (2) the effective ways of visualization for the opposition between direct and figurative meanings of lexical items; (3) the problem of the borderlines between semantic fields, which seem to be very subtle. These problems are discussed in detail in the paper, as well as possible theoretical decisions and semantic modelling techniques that could overcome these bottlenecks.

The Optative negative of Beja is a multifunctional paradigm which encodes optative, hortative and jussive grammatical meanings, depending on the person, as well as participant-internal and participant-external modalities of impossibility and unnecessity. It is also the sole paradigm licensed in balanced embedded clauses. Based on a large corpus of naturalistic first-hand data, this study analyses the various uses of the paradigm, provides an account of its evolution from the pre-modal stage to the post-modal stage on the basis of language internal morpho-syntactic cues, and proposes a diachronic semantic map, based on van der Auwera and Plungian (1998) model. It shows that semantic maps are not only useful for typological purposes, but also for language internal studies, helping understand the semantic shifts that occurred in the grammar of unwritten languages with no recorded history.