{[['
']]}
']]}
But while lexical semantics focuses on content words, such words cannot be studied in an agrammatical vacuum. Some lexical properties, like Aktionsart (lexical aspect) have effects throughout the sentence. So, for instance, a difference between the verbs spot and see can be described in terms of aspectual properties of the verbs: spot describes a punctual event, while see does not. This in turn affects which tense and aspect markers can be present in the same clause and how such markers are interpreted. So, I saw the bird all day long can describe a continuous seeing event, while I spotted the bird all day long must be interpreted as repeated instances of spotting events. Because of the effects of the verbs’ semantics on other elements in the sentence, Aktionsart (and other topics, like thematic role assignment) is often presented as ‘semantics’ in textbooks, rather than as ‘lexical semantics’. This can create problems both for students’ understanding of lexical semantics and for instructors’ planning of a lexical semantics course that complements a general semantics course. The approach taken below is to offer a broad range of lexical semantic topics.
2.1 Topics in lexical semantics
Lexical semantics fits into linguistics curricula in various ways. Some of the most common ways are:- as a sub-module in a semantics course (often lower-mid level in degree)
- as part of a course on vocabulary/lexicology—including morphology, etymology, lexicography as well as semantics (often lower-mid level)
- as a free-standing course (often upper level)
- Topics 1-4 (basic issues) should be covered in any linguistics degree, and may be covered in courses other than lexical semantics.
- Topics 3-4 and 6 are covered in most short lexical semantics modules (e.g. within a general semantics course).
- Coverage of topics 5 and 7-9 is more uneven across departments, but at least some aspects of these would normally be covered in an upper-level lexical semantics course.
| General Topic | Basic issues to discuss and useful resources | Taking it further |
|---|---|---|
| What is a lexicon? |
|
|
| What is a word? |
|
|
| What is meaning? |
|
|
| Meaning components: basics |
|
|
| Alternatives to ‘Classical Theory’ | Some selection of the following (covered briefly in Löbner 2002, Saeed 2003—more specific textbooks listed below):
|
Primary theory sources:
|
| Semantic relations |
|
|
| Topics in verb meaning |
|
|
| Topics in noun meaning | [basic issues generally raised under topics 4/5] |
|
| Topics in adjective meaning |
|
2.2 Approaching lexical semantics from other angles
Stand-alone lexical semantics courses are fairly unusual, but aspects of lexical semantics can be taught within a range of other disciplinary and cross-disciplinary courses. As already mentioned, this most commonly happens in general semantics courses, and can also be part of lexicography or lexicology courses. The latter, under various names, have become more common lately, particularly as introductory linguistics courses—as words provide an accessible gateway to many linguistic concepts.Lexical semantics courses can incorporate other (inter-)disciplinary interests. Courses in other (sub-)disciplines, including the following, can benefit from inclusion of some lexical semantic topics:
Pragmatics – No semantics course can help but to tread on the toes of pragmatics, and some theoretical approaches (particularly Cognitive Linguistics) have done away with the distinction between semantic and pragmatic competence. Still, one of the first challenges in learning about lexical semantics is to be able to make the distinction between a word’s contribution to the meaning of an utterance and the contributions of context (pragmatics) and co-text (the phrasal context). Pragmatic accounts have been proposed for many lexical semantic issues, such as polysemy (e.g. Nunberg 1979, Blutner 1998) and semantic relations (Murphy 2003).
Morphology – Just as there are many interfaces between syntax and sentential semantics, so there are between morphology and lexical semantics. One is the question of whether word class is semantically determined (see Table 1, topic 2). The semantics of derivational morphemes and derived words also provides fertile thinking ground. Kreidler’s Introducing English semantics (1998) has an accessible chapter for beginners, while Lieber (2004) provides a theoretical account that fits with Jackendoff’s Conceptual Semantics.
Psycholinguistics – Most lexical semantic issues can be addressed from a psycholinguistic perspective, and psycholinguistic methods offer evidence concerning how words and meanings are organised in the mind. Aitchison (2002) provides an introduction to many of these issues. Reeves et al. (1998) outline some of the experimental methods used in investigating the mental lexicon, some of which can be replicated in the classroom.
Language acquisition – Unlike grammar, vocabulary is acquired throughout life, so some of the issues in lexical acquisition can be addressed from an adult first- or second-language angle. An appreciation for some basic issues in semantics—such as the arbtrariness of the sign and how we solve Quine’s (1960) gavagai problem (discerning the meanings of words in context) can be gained through consideration of how one initially discovers that bits of sound have meaning. Most child language textbooks include sections on lexical acquisition. Clark (1993) and Bloom (2000) take particular perspectives on some of these issues. In second language acquisition, Carter (1998) introduces many of the relevant issues, and the recent influence of lexical teaching/learning strategies provides another point of engagement with the lexicon.
Anthropological linguistics, field linguistics, typology – Cross-linguistic lexical comparison has a long history in anthropology, particularly with reference to kinship terms, biological classification and colour (Berlin and Kay 1969). Lexical-semantic typology (e.g. Talmy 1985) and contrastive lexical semantics (e.g. Weigand 1998) raise issues that are relevant to language courses as well. Linguistic relativity, the idea that the language one knows can affect one’s means of thinking, is enjoying some reconsideration (e.g. Gumperz & Levinson 1996). Some of the more promising avenues for such study involve morphological categories, such as gender. The converse idea, that lexis is affected by culture, has been explored in depth in Anna Wierzbicka’s work (particularly 1992 and 1997). These issues are usually very popular with students.
Computational linguistics – Much lexical semantic work nowadays is done in computational linguistics/natural language processing (NLP), including much work on polysemy/ambiguity resolution and the development of semantic networks. WordNet (Fellbaum 1998) has enjoyed particular success as an NLP tool and model of lexical structure. There is a large body of work on ambiguity resolution, and Pustejovsky’s (1995) Generative Lexicon Theory (see below) has arisen though and been applied to such NLP concerns.
2.3 Lexical semantic theories
Whether to teach a particular theoretical perspective on lexical semantics is a tricky question, since in a sense, there’s no such thing as a free-standing ‘lexical semantic theory’, but instead there are semantic theories that pay more or less attention to representation at the lexical level. If one were to try to cover the topics listed in Table 1 with reference to a single theory, it is unlikely that even coverage of the topics would be possible, since different theoretical perspectives have taken different starting points.For first-year courses, a pre-theoretical approach is usually most attractive. This can be done, for instance, by taking a lexicographical approach—i.e. what are the issues that one needs to understand in order to write dictionary entries? Jackson’s textbook Words and their meaning (1988) took this kind of approach. Hudson’s workbook Word meaning (1995) introduces basic lexical semantic concepts with minimal theoretical discussion, and Cruse’s Meaning in language (2000b) is heavily focused on lexical issues. In addition, there are lexicology textbooks (Jackson and Zé Amvela 2000, Singleton 2000) that are suitable for lower-level courses. These are generally more suitable for courses that intend an ‘introduction to linguistics through words’, since their treatments of basic semantic issues, such as polysemy, are rather cursory.
In more advanced courses, theoretical approaches are appropriate, but which ones should be tackled is to some degree a matter of taste and of the rest of the curriculum—in that it could be confusing to teach a theory for lexical semantics that is incompatible with the theories taught in the department for general semantics and grammar. That said, while general semantics courses are often taught from a formal, model-theoretic perspective, lexical semantics teaching is generally approached from non-formal stances.
My own preference is to contrast different theoretical approaches to a particular topic (without presenting formalisms that require knowledge of particular logical languages)—but this presents particular challenges. First, the approaches to be contrasted may have very different basic assumptions about the nature of meaning or the lexicon, and thus a fair amount of background to the theories must be presented. Second, they may not be trying to answer exactly the same questions or to cover the same set of phenomena.
There are at least three ways of getting around these issues. One is to present from the beginning two general types of approach (e.g., componential vs image-schematic), examining in detail first their basic assumptions, and then investigating how the contrasting theories have (or would) approach particular phenomena. We can call this the ‘comparative approach’. Because one is dealing with ‘families’ of theories in this way, it can be possible to examine a range of problems from contrasting perspectives. Still, it does require some theoretical juggling. (This is the approach of Murphy, forthcoming.) Another possibility is to take an ‘envelope-pushing’ approach, in which a particular theory is taken as a starting point, but then challenged to account for issues that have been discussed from purely descriptive or other theoretical approaches. This is most suitable to upper-level courses, in which the students are ready to develop their own theoretical accounts. The final option is to simply teach a theory as it is presented in its fundamental literature. However, while this last option may be appealingly straightforward, it is the least promising for instilling critical thinking skills or a broad perspective on the field. The comparative and envelope-pushing approaches can also be valuable for inspiring further research projects, for example in a final-year dissertation.
One of the greatest challenges for teaching semantics from a theoretical standpoint is the lack of theory-specific textbooks for the undergraduate level. The only current texts in this vein that I am aware of are Goddard’s Semantic Analysis (1998), which introduces mostly lexical issues in the Natural Semantic Metalanguage framework, and Cognitive Linguistics texts (most of which are not limited to lexical semantic issues). Some general semantics textbooks (e.g. Löbner 2002, Saeed 2003) give comparative overviews of some theoretical approaches, but for the most part, it can be difficult to find textbooks that are suitable to non-formal theories. Frawley’s Linguistic Semantics (1992) covers many issues of lexical interest at some level of comparative theoretical detail. Another option is to develop a reading pack, based on shorter primary materials, including book chapters, journal articles and encyclopedic overviews.
Some relevant, mostly current theoretical approaches are listed below, with citation of their foundational literature or (where possible) textbooks—marked here by T. In general, they can be divided into two types: componential and schematic. Componential approaches rely on a language-like system of meaning representation involving a limited number of primitive symbols in some kind of grammar—the classic example being the model in Katz and Fodor 1963. Schematic approaches take the position that word meaning must be approached within more complex conceptual structures, often relying on a more image-driven representational form, such as the ‘image schemas’ of Lakoff 1987. Componential approaches are more generally associated with the goals of generative linguistics, and schematic ones with cognitive linguistics, although there is a wide range of variation among all these approaches.
| Theory | Characteristics | Starting point references |
|---|---|---|
| Cognitive Semantics |
|
|
| Conceptual Semantics |
|
|
| Frame Semantics |
|
|
| Generative Lexicon |
|
|
| Natural Semantic Metalanguage |
|
3. Teaching through student-led research
Words provide self-contained packets of language about which many types of investigation can be carried out. Teaching lexical semantics is thus particularly exciting for the opportunities that it allows for student-led, original research.3.1 Research methods for lexical semantics
Students can research word meaning using a variety of tools, including introspection, fieldwork, dictionaries, corpora and (where appropriate) psycholinguistic experimentation, as discussed in turn below.Introspection: Asking oneself how one uses language is the classic linguistic method, and it should be used throughout a lexical semantics course. Beginning students usually need some guidance to bring them from “what does this word mean to me?” (a question whose answer may not be reliable) to finding answers for the question “how can I test the semantic properties of this word?” For students whose native language is not the language of the course, introspection is usually not a suitable main means of discovery. I therefore allow students to treat any assignment that calls for introspection as also allowing ‘field methods’.
Field methods: Where native intuitions are not available or where they may vary, one may instead (or in addition) quiz native-speaker informants about the acceptability of a word in various contexts (or about the boundaries of the word’s sense, etc.). Here, a little more training is needed, in part to avoid wasting informants’ time and patience. This can be achieved by modelling field methods practice in the classroom and then asking the students to devise a list of questions/examples that they will ask their informants about. For any assignment that is the same for all students on the course, I enforce a rule that their informants cannot be members of the class, nor other Linguistics students, so that the ‘field methods’ cannot be mistaken for ‘collusion’.
Dictionaries: Dictionary definitions can provide a good starting point for thinking about a word’s meaning, the nature of polysemy and the relation between descriptive and prescriptive attitudes to language. Advanced learners’ dictionaries often provide more ‘grammatical’ information about words, including information about collocations and more specific grammatical categories (e.g. count/mass nouns), which can be valuable for both non-native and native speakers. The Oxford English Dictionary, on the other hand, provides plenty of etymological information and examples of usage. Both types can be valuable for different kinds of activities. Some activities using dictionaries include: using a number of dictionaries to map the sense boundaries of a particular word, comparing actual uses of words to their dictionary definitions (are their senses more fluid than the dictionary records?), and determining the principles underlying the organisation of information in a thesaurus.
Corpora: Corpus linguistics offers a means to supplement and/or challenge introspective evidence. The problem for any lexical semantics course, however, is whether the facilities and time are available to provide students with access to a corpus, corpus software and the requisite skills to use them. For my own courses, the answer has been ‘no’, but it is still possible to introduce corpus analysis. A basic way is to provide students with printouts or word-processor/pdf files of pre-prepared concordances. Once students know a bit about corpora, they can have a hand in designing the methodology for a particular corpus investigation, with the tutor or technician still charged with executing it. If the question to be put to the corpus is simply “can one say X in English?”, it will usually suffice to have students search for exact phrases on the internet using a standard search engine.
Experiments: Most lexical semantics courses will not have the time/facilities to teach experimental methods, nor to teach students to use the types of software usually used in psycholinguistic experimentation. Some experiments, however, like some used by Eleanor Rosch (1978) to demonstrate prototype effects, can be carried out with pen and paper and extended to different words/categories.
3.2 Structuring a course around original research
Stand-alone assignments and seminar activities can be based on any (or several) of the above methodologies. It is also possible to devise a course in which student-led research provides the themes to be discussed.Adopt-a-Word: My own lexical semantic courses are structured around an assignment called ‘Adopt-a-Word’, which allows students to apply a number of key concepts and methodologies to a single word, thus building up a portfolio of short assignments around the theme of a word. (The Adopt-a-Word scheme for a year 1 lexicology course is described in more detail at the LLAS event report, listed in the web links below.) That course is also designed to establish key disciplinary, academic and transferable skills. (See Hudson 2003 for discussion of skills development.) For a more advanced/specific lexical semantics course, suitable assignments can be designed. (Murphy, forthcoming, provides a number of ideas, most of which have been tested in a final-year course in the US.)
In my courses, students receive at the beginning of the term a document with a dozen or so possible Adopt-a-Word assignments. After each student is assigned (or chooses) their own word, it is up to the student to determine (with some guidance) which assignments suit the word. For instance, an assignment on argument structure may not be suitable for many nouns. The student then does a number of assignments on the same word. The same assignments can be used without adopting a word—i.e. the tutor could assign different words for each assignment, thus allowing students to be assessed on every element of the course, rather than just those that suit their word.
One advantage of the Adopt-a-Word method is that the student becomes an ‘expert’ on a word, and thus later assignments are informed by work for previous ones. Another is that using the adopted word as a theme for the term allows for greater continuity and coherence in a survey course that flits from descriptive/theoretical topic to topic. Finally, involving students in original research greatly reduces the potential for (intentional or unintentional) plagiarism in essay-writing.
Dissertations in lexical semantics: While Adopt-a-Word allows for a broad-based approach to semantics/lexicology, some upper-level courses require more in-depth research projects, as can be assessed through a dissertation. Here, again, it is very doable for students to tackle original lexical semantic research, preferably using more than one of the methodologies discussed above. To give some examples, here are short descriptions of some recent final-year dissertations at Sussex:
Is David Beckham black? An investigation of the meaning of the racial term black in the UK, testing hypotheses based on Social Identity Theory posited in Murphy 1997. Data sources included dictionary definitions, questionnaire responses and examples of usage from the press.
Child and adult in Japan and Britain. A study by a Japanese native speaker based on Wierzbicka’s (1997) contention that the meaning/use of particular words can offer insight into cultural differences. The student used dictionary definitions, legal definitions and questionnaire responses to contrast the denotations and connotations of words for ‘child’ and ‘adult’ in the two cultures.
Bizarre: a case study of near-synonymy. This student used corpus data and affect-based questionnaires both to demonstrate that exact synonyms for bizarre do not exist and to determine the ‘semantic prosody’ (Louw 1993) of bizarre and its synonyms.
4. Closing note
This guide has briefly raised some issues concerning the teaching of lexical semantics, but it has necessarily been based mostly on my own experience. In order to improve the guide, please contact me (m.l.murphy@sussex.ac.uk) with information on your successes in teaching lexical semantics and/or feedback on your experience of any of the methods discussed here.Sumber https://www.llas.ac.uk/resources/gpg/2821
Posting Komentar