This feature allows to generate a graphic representation of the supposed genetic relationships between the language set included in the database, in the form of a genealogical tree (it is also implemented in the StarLing software). The tree picture also includes separation dates for various languages, calculated through standardized glottochronological techniques; additionally, a lexicostatistical matrix of cognate percentages can be produced if asked for.
The tree can be generated by a variety of methods, and you can modify some of the parameters to test various strategies of language classification. The pictures can be saved in different graphic formats and used for presentation or any other purposes.
This option displays the full description for the selected database, including: (a) the complete list of primary and secondary bibliographical sources for the included languages, including brief descriptions of all titles; (b) general notes on said languages, e. g. sociolinguistic information, degree of reliability of sources, notes on grammatical and lexical peculiarities of the languages that may be relevant for the compilation of the lists, etc.; (c) details on the transcription system that was used in the original data sources and its differences from the UTS (Unified Transcription System) transliteration.
This option, when checked, uses a set of different color markers to highlight groups of phonetically similar words in different languages with the same Swadesh meaning.
Phonetic similarity between two different forms is defined in the GLD as a situation in which the aligned consonants of the compared forms (usually the first two) are deemed «similar» to each other. In order for two consonants to be «similar», they have to belong to the same «consonantal class», i.e. a group of sounds that share the same place and a similar manner of articulation. The current grouping of sounds into sound classes can be found here.
Accordingly, the aligned forms undergo a process of «vowel extraction» (all vowels are formally assumed to belong to «class H», together with «weak» laryngeal phonemes), and the individual consonants are then converted to classes, e. g. dog → TK, drink → TRNK (in comparisons, only the first two consonants will be used, so, actually TR), eat → HT (word-initial vowel is equated with lack of consonant or «weak» consonant), fly → PR (l and r belong to the same class) and so on.
If both of the first two consonants of the compared forms are found to correlate, i.e. belong to the same class, the words are deemed similar (e. g. English fly and German fliegen both have the consonantal skeleton PR). If at least one differs, the words are not deemed similar (e. g. English tooth → TT and Old Norse tɔnn → TN, although they are etymological cognates, will not pass the similarity tense because of the second position).
In most cases, checking this option will highlight phonetically similar forms that are also etymological cognates and share the same numeric cognation indexes. Occasionally, however, the checking will also yield «false positives» (accidentally phonetically similar forms that do not share a common origin) and «false negatives» (phonetically dissimilar forms, not highlighted, but actually cognate). It should be noted that one should never expect this method to yield a 100% accurate picture of etymological cognacy. Rather, the method is useful for the following goals: (a) assess the amount of phonetic change that took place between related languages; (b) give a general idea of the degree of closeness of relationship for those languages where phonetic correspondences have not yet been properly established; (c) assess the average number of «chance similarities» that may arise between different languages.
The last task is particularly instructive if the «Highlight...» option is used between two different languages from different databases, i.e. not related to each other or distantly related: in most cases, it will yield around 2-3 accidental color highlights, but occasionally, the count may go as high up as 5 or 6.
This option unfolds all of the notes that accompany the individual forms in the database. Sometimes these notes only consist of a basic reference to the bibliographical source, but at other times, they can be quite expansive, which makes browsing through the wordlist quite cumbersome. By default, the notes stay hidden (each note can also be opened separately by clicking on the sign next to the word).
Brassett et al. 2006: 206. Cf. also: ɕi̯ɛ21-ʆi24 'all' [ibid.], ɕiŋ24-kʰi55 'all' [ibid.], yi24-si55 'all' [Brassett et al. 2006: 49, 207]. The difference between these quantitative pronouns is not quite clear. Additionally, the particle tu̯ɔ21 is used with interrogative pronouns to form meanings like 'anyone', 'anything' [Brassett et al. 2006: 49].
Brassett et al. 2006: 168. The main part of the word is the same as 'skin' q.v. Also attested in an extended variant: kʰa21=mɔŋ21=tʰa24-pʰa21, where kʰa21=mɔŋ21 = 'tree' q.v.
Brassett et al. 2006: 156. Same word as 'to eat' q.v.; dubious, since the word is only attested in one example in this meaning (but a good one: "I was bitten by a dog").
Brassett et al. 2006: 186. This is probably the transitive verb, opposed to the intransitive ta21 [Brassett et al. 2006: 165] (based on the location of these items in Brassett's classified vocabulary). Cf. also wu24 'to burn' [Brassett et al. 2006: 187].
Tian 1986: 175. The actual attested form is even 'longer', referring specifically to 'finger-nail': ɕe35=mi55=tʰi=21=kʰo53-tʰa53, where the first three syllables = 'finger'.
Cf. also Longshan me35=lan53-uen21 [Zhang Weiquan 2006: 364]; Longshan mɯe35=la55-ɣoŋ21, Luxi yin21-dɯ21-bo33 [Chen 2006: 176]. The meaning 'cloud' in Northern Tujia is always expressed by the bisyllabic compound *la1-ɣɔŋ3, whose individual parts are semantically obscure. In all known Southern Tujia dialects, the original word has been replaced by Chinese borrowings.
Cf. also Longshan xa53-le21 [Zhang Weiquan 2006: 78]; Longshan xa55-lie21, Luxi cʰie55 [Chen 2006: 180]. There is no certified Common Tujia equivalent: all Northern dialects have the disyllabic formation *xa5a-li̯e3 (individual meanings of the two syllables are unknown), and all Southern dialects have *cʰe5a. It is quite possible, although not a certainty, that the Northern forms really go back to a monosyllabic borrowing from a Hmong source (if North Tujia *xali̯ɛ < *xle < *ɬe, cf. Qiandong Hmong ɬa, etc.). If so, the Southern forms may be more archaic.
Build treei