Y. Ferland

DRDC Valcartier, Information and Knowledge Management, Quebec, Canada


Alike traditional atlases and gazetteers, toponymic databases are normally structured as tables presenting, for each unique place-name entry, a double binomial of : name designation {generic, specific} (or the reverse in some grammars of languages) and place location {latitude, longitude} in the numeral form of coordinates of its cartographic centroid. The generic designates the geographical feature in the official language(s) of the country, and the specific is the proper name of any particular place or feature that is located at these coordinates. According to the rationale of official (i.e. administrative) standardization of toponymy, one pretends that this simple nomenclature is the most universal common case for general convenient use, and also well-suited for geospatial database models (computer tabular memory, not transactional).


But it is certainly not appropriate for many usages of those kinds of data, as for: map-making, information retrieval, social science research, and strategic planning. Problems for gazetteer users are not provoked, but amplified by new information technologies that impose a double constraint with respect to administrative standards. Effectively, exclusivity of name entry considers all other names as variants (if at least doing so), whatever their intrinsic value, and name placing is automatically fixed at centroid position without respect to entity spatial coverage.


Specialized users of such databases may be classified into four groups : cartographers replacing toponyms at their right significant location on maps with respect to entity geometry (representational data structure), users consulting the list for spotting the place named on the map, public managers who control any place within their territorial jurisdiction, people who searches information to complement by the means of those place-names.


For defence, intelligence, and security purposes, even governemental users of any kinds can hardly be satisfied with present unique names having a standardized, administratively correct spelling. Are also valuable: vernacular, minority, or exogenous names, linguistic or historic variants, historical or cultural displacement of named entity coverage, linguistic evolution of feature identification, false or peculiar generic not concording with the present feature, geospatial term included in specific… agents may need them all. In military missions abroad, readability of place-name meaning is essential but it disappears with direct translation or transliteration from local languages. In many languages, important grammatical aspects as gender, articles, or ellipsis are just absent of gazetteer tables.


To disambiguate, terminological and technological innovations are considered since geospatial intelligence experimentation, for developing smart transactional functions to adapt semantic usages of toponyms.