next up previous contents
Next: The LU node Up: An overview of the Previous: Linguistic levels

GPMU and CGPMU

GPMU

The lexical architecture does not impose any hierarchy among the GPMUs. Differenciations between long forms, acronyms, abbreviations, variations, etc. are expressed by features. Also, preferences motivated by standardization or by pragmatic reasons can be attached to a GPMU without constraining the overall model. For example, "DNA" and "deoxyribonucleic acid" are two forms of the same meaning. They are described by two GPMUs each one with the appropriate feature indicating their type "initials" or "long form".

Variations like "cheque" and "check" are also dealt with by creating two GPMUs. In order to specify that the first orthography is British and the second is American, pragmatic information concerning the geographic usage is attached to each GPMU as a whole.

The GPMU can be either seen as the actual receptacle in which the orthographic, phonological and morphological information is stored or the node from which these constituents can be accessed with specific identifiers. This alternative does not change the role of the GPMU as the only surface object directly known to the LU. But mainly for facilitating implementation it has been decided that the GPMU constituents could be separated. The reason is that it may be easier to manipulate small objects than big ones. Also, this way, redundancy is dealt with straightforwardly.

CGPMU

Complex words - compounds, phrases, expressions, etc. - are considered similar to simple words with regard to the two node organization. The word "Application Programming Interface" is described by one LU for the single meaning and two GPMUs ; one for the long form and one for the initials "API". The difference lies in the treatment of the long form, at the GPMU level.

In order to account for the internal structure of complex words, a special kind of GPMU, the CGPMU (Complex GPMU) has been created. This object is similar to the GPMU in many respects. It contains orthographic, phonological and morphological information but, in addition, it refers to the components - GPMUs of simple constituents - of the complex word, gives its morphological structure and global behaviour and expresses constraints on the components, if required.
For example, the CGPMU for "Application Programming Interface" refers to the three GPMUs "application", "programming" and "interface", and it also specifies that the mark of the plural only apply to "interface".
The content of a CGPMU can be summerized as follows:

The CGPMU is an extension to the core MULTILEX standard. It was developed within a MULTILEX-EUROLANG gif joint group.



next up previous contents
Next: The LU node Up: An overview of the Previous: Linguistic levels