4. RESULTADOS EXPERIMENTALES
4.2.2. Resultados de los ensayos a macroescala de 4 años (Bailén)… 83
4.2.2.2. Caracterización mineralógica de las columnas de arcilla…. 88
Thesaurus construction is guided by international and national standards. National standards are constructed in accordance with international standards (ISO 5964-1985, 1). The thesaurus construction standards stipulate the use of three basic inter-term relationships, which are the equivalence relationship, the hierarchical relationship and the associative relationship. (Aitchison & al. 1997, 47; ISO 2788-1986, 13) The relationships are regarded as language-independent and also culture-independent, and therefore they apply equally to both monolingual and multilingual thesauri (ISO 5964-1985, 25, see more in Hudón 2001, 74-75).
Equivalence relationship (author note: as a basic relationship within one language in a thesaurus – not to be mixed here with translation equivalence between different languages) refers to synonyms and quasi-synonyms. In indexing all the members of an
equivalence set are regarded as referring to the same concept. One of the terms in the set is designed as the preferred term, and the other(s) as non-preferred. When all the languages in a multilingual thesaurus have equal status, then every preferred term in one language should be matched by a corresponding term in the other language(s). Since languages frequently differ in the number of synonyms which refer to a given concept, an exception for one-to-one demand are non-preferred terms, for which it is not necessary to establish correspondences in different languages. (ISO 5964-1985, 26)
Hierarchical relationships cover two different kinds of logical relationships – the generic relationship and the hierarchical whole-part relationship, which applies only to certain classes of terms, such as geographical entities and parts of the body. It is the basic relationship which most obviously distinguishes a systematic thesaurus from other organised lists of terms. (ISO 5964, 26)
ISO 5964-1985 sees, as a general rule, that any hierarchy which the users of one language regard as logically acceptable should appear to be equally valid when its terms have been translated into another language (ibid, 26). Hierarchical relationships can cause translation problems – or reveal them:
“It can further be asserted that if a hierarchy is accepted as “natural” by the users of one language, and this hierarchy is not, for cultural reasons, regarded as equally valid by the users of another language, then it may not be possible to achieve acceptable equivalences at the level of the individual terms. Such fundamental differences between the categorical systems of different language users would tend to indicate that the terms in these languages either refer to different concepts, or they express the same basic concepts from such different viewpoints that the hierarchy expressed in the source language cannot be translated, as it stands, into the target language. Faced with this situation, the term which appears to belong to different categories, depending upon the language or culture, should be treated as polyhierarchical term (---) : that is, it should be admitted as a subordinate term in each of these different hierarchies, and both hierarchies should then be translated into the other language.”
“Translation of a hierarchy from source to target language will sometimes reveal a gap in a hierarchy in one of the languages. If the target language already possesses an acceptable equivalent to the “extra” term in the source language, and provided that this term is regarded as useful by either of the language communities, then the “missing”
term should be added to the hierarchy in the target language.” (Ibid, 26)
“- If an acceptable equivalent does not exist, this should be treated in accordance with the recommendations given about equivalence degrees and their solutions” (ibid, 27).
The associative relationship is considered to be the hardest one to define (ibid, 27):
“It refers to the link between two terms which do not form an equivalence set, nor can they be organized as a hierarchy, yet they are associated in common usage to such an extent that the user who refers to one of these terms in an index or thesaurus should be directed to the other term. It is necessary to exercise a measure of control over terms linked by this relationship. It can be stipulated, for example, that one of the terms should form a necessary component in any definition or explanation of the other.”
(Ibid, 27)
According to ISO 5964-1985 associative relationships in multilingual thesauri should be represented one-to-one between different language versions, and due to cultural reasons, there should be generosity in providing them, but at the same time they should not serve the needs of one (cultural) group.
“Before the associative relationship which has been recognized in one language is transferred to another, it should be examined to determine the extent of its validity. If it appears to apply to only one group of language users, it should generally be excluded. In that case, the terms should also be re-considered to ensure that they do, in fact, refer to the same concept. Despite this injunction, a multilingual thesaurus should usually contain a richer variety of associative relationships than a monolingual thesaurus in the same field, since it will benefit from the viewpoints of different language users.” (Ibid, 27)
When selecting or evaluating preferred terms to a thesaurus it is noteworthy, that according to the SFS 5471-1988 the selected descriptor should be as clear, short and unambiguous as possible, and its meaning must be the same as that generally used in the field of thesaurus. When focus on multilingualism the guidelines towards loan words and translation of loan words is also interesting. The ISO 2788-1986 states:
“Terms from other languages are sometimes encountered as “loan words”, i.e. foreign terms which are accepted as newly-coined terms. If these terms are well-established, they should be incorporated into the thesaurus. Occasionally a loan word and a putative translation co-exist. If the loan word is more widely accepted, it should be treated as the preferred term, but if the translation becomes well-established, this should be preferred. Reciprocal references should be made between the preferred and non-preferred terms.
Example:
X-RAYS
UF roentgen rays
Roentgen rays USE X-RAYS” (Ibid, 7-8)
Here, one should also keep in mind, even in the case of a citation loan that the same concept must exist in both languages. YSA (1.4.2009), for example, includes terms such as aerobic, afasia (aphasia), balalaika, bingo, bulimia, calvados, rock, punk, reggae etc since they are established concepts and words in Finnish common and/or specific language. Sometimes the citation loans are adapted into the Finnish language system by different transliteration, (e.g. cyberpunk USE kyberpunk). In translation science is often repeated, that these kinds of terms may represent false friends – terms which look the same but have different meanings.
In the guidelines (see e.g. ISO 2788-1986, 30) it is stated that the correctness of terms should be verified before they are admitted to thesaurus:
“The following types of authority should be checked when terms are accepted as candidates for inclusion:
standard technical dictionaries and encyclopaedias;
existing thesauri;
classification schemes.
Subject specialist, especially those with some knowledge of indexing and documentation, may also be consulted.” (Ibid.)
Equivalence relationship is according to ISO 2788-1986
“--- relationship between preferred and non-preferred terms where two or more terms are regarded, for indexing purposes, as referring to the same concept. In a printed thesaurus, preferred and non-preferred terms should be distinguished typographically whenever possible. Reciprocity is expressed by the following conventions:
USE, written as a prefix to the preferred term;
UF (use for), written as prefix to the non-preferred term. (Ibid, 13)
The general equivalence relationship covers two types of terms, namely synonyms and quasi-synonyms. Terms considered as synonymous are those, “whose meanings can be regarded as the same in a wide range of contexts, so that they are virtually interchangeable”. They are more common in a controlled indexing language, where meanings are more restricted, on purpose, than in the natural language. In practice, there are various types of synonyms: terms of different origin (e.g. polyglot vs. multilingual);
popular names and scientific names (aspirin vs. acetylsalicylic acid); common nouns and trade names (vacuums flasks vs. thermos flasks); variant names for emergent concepts (hovercraft vs. air cushion vehicles); current or favoured terms versus outdated or deprecated terms (developing countries vs. underdeveloped countries); variant spellings (geese vs. goose); terms originating from different cultures sharing a common language (lifts vs. elevators); abbreviations and full names (PVC vs. polyvinyl chloride); and the factored and unfactored form of a compound term (coal + mining vs. coal mining). In the case of synonym, preferred terms should be selected to serve the needs of the majority of users, bearing in mind the general guidelines for indexing term (form) selection. Predictability should be taken into account, and the selected principle (e.g.
popular names rather than scientific names) should be used consistently. (Ibid, 14) Quasi-synonyms are defined as “terms whose meanings are generally regarded as different in ordinary usage, but they are treated as though they are synonyms for indexing purposes” (ibid, 14). It is also pointed out, that quasi-synonyms are likely to be least common in an indexing language, which covers several disciplines, and terms should be treated as quasi-synonyms only in fringe subject areas –they should not be used as a means for reducing the number of preferred terms in an indexing language.
(Ibid, 14)
In the context of equivalence relationships upward posting as a third case is discussed. Upward posting “refers to a technique in which the name of a class and also the names of its members are treated as an equivalence set, the broader term then functioning as the preferred term”. This technique is sometimes used to reduce the number of terms in an indexing language, but it should generally be avoided. “If employed, it should be applied only to terms in the fringe area of the subject field covered by the thesaurus”. (Ibid, 14-15)
According to ISO 2788-1986 -standard hierarchical relationship is the one, that most distinguishes a systematic thesaurus from an unstructured list of terms, such as
those from a glossary or dictionary. “It is based on degrees or levels of superordination and subordination, where the superordinate term represents a class or whole, and subordinate terms refer to its members or parts”. It is expressed with the abbreviation BT (i.e. broader term), written as a prefix to the superordinate term and NT (i.e.
narrower term), written as prefix to the superordinate term. The hierarchical relationships cover three logically different situations – the generic, the hierarchical whole-part and the instance relationship. It should be pointed out, that every subordinate term should refer to the same basic kind of concept as its superordinate term, i.e. both the broader and narrower term should represent a thing, or an action, or a property, etc.
(Ibid, 15)
ISO 2788-1986 states, that associative relationship is easier to define in terms of negative rather than positive characteristics. Furthermore,
“It covers relationships between pairs of terms which are not members of an equivalence set, or can be they be organized as a hierarchy in which on term is subordinated to another, yet they are mentally associated to such an extent that the link between them should be made explicit in the thesaurus, on the grounds that it would reveal alternative terms which might be used for indexing or retrieval. This relationships is reciprocal, and is indicated by the abbreviation “RT” (related term), or its equivalent in other languages. (Ibid, 17)
ISO standard (ibid) puts stress on the rule that thesaurus constructors should exercise strict control over the choice of terms linked in this way, and avoid subjective judgements.
“As a general guideline it can be stated that one of the terms should be strongly implied, according to the frames of reference shared by the users of an index, whenever the other is employed as an indexing term. More specifically, it will frequently be found that one of the terms is a necessary component in any explanation or definition of the other, to the extent that the term “birds”, for example, forms a necessary part of the explanation of “ornithology”. (Ibid, 17)
Using associative relationship, two kinds of terms can be linked - those that belong to the same category and those belonging to different categories. Terms belonging to the same category “relates to siblings with overlapping meanings, such as “ships” and
“boats”, where each of the terms is amenable to an exact definition, (consequently they do not form an equivalence set), yet they are sometimes used loosely and almost interchangeable, so that the user seeking documents on one of the terms should be reminded of the other. Terms belonging to different categories are of various kinds, but they should satisfy the requirements that one of the terms should be strongly implied by the other. As representative examples of typical relational situations encountered in practice is offered: a discipline or a field of study and objects or phenomena studied (forestry RT forests); an operation or process and its agent or instrument (temperature control RT thermostats); an action and the product of the action (weaving RT cloth); an action and its patient (harvesting RT crops); concepts related to their properties (poisons RT toxicity); concepts related to their origins (Dutch RT Netherlands); concepts linked by causal dependence (bereavement RT death); a thing and its counter agent (plants RT herbicides); a concept and its unit of measurement (electric current RT ampere); and
syncategorematic phrases and their embedded nound (model ships RT ships). (Ibid, 17-19)
Not only do hierarchical and equivalence relationships give valuable information especially from the translatability’s and (sub)cultural context’s point of view but also the associative relationships provide valuable information about the informants’ conception of the studied concepts; they are thus also in focus but mainly as a contextual and semantic information. Recognition of associative relationships may be very cultural- and group-bound action, and that is the reason why the creation of an associative thesaurus (Term Association Thesauri, TA-Thesauri) is considered to be especially demanding, although worth the effort (see e.g. Iivonen1989, 176). Associative relationships are important (Kristensen & Järvelin 1990, 15) or equally or more important than hierarchic relationships (Iivonen 1989, 176-178; Tudhope et al. 2001) for expansion of the query and improving recall. Furthermore, as Hudón (2001) has stated,
“A critical function of the thesaurus is that of helping users in making sense. The defining function of thesaurus relationships is well established: Relationships “define,”
admittedly not always very clearly, by providing a context of sorts that determines the place of a concept in its semantic environment, and by supplying more or less explicit information on the intension and extension of this concept. Relationships also serve as navigational aids to arrive at the intended destination or at the most interesting destination on a semantic map. In a multilingual context, both the defining and navigational functions of thesaural relations acquire even more importance.” (Ibid, 69) Since related terms are recommended for use in the expansion of the query and improving recall and also often seen as culturally biased, they can give very valuable information about the content and context of the terms in a specific context and are thus also very important for the study of multilingual thesaurus construction. It is of importance to see how they are represented within the search results of this study, that is, in the selected general and social science thesauri (about material see esp. 6.3.2 Thesauri and about results esp. 9.1 Semantic lexical networks in thesauri). The hierarchical relationships also provide similar kind of information; but otherwise, the focus is on equivalence relation (due to focus of the study).