Documentary and descriptive linguists often produce dictionaries of understudied languages, and speakers of these languages often show great interest in these dictionaries – typically a lot more interest than in grammatical analyses, which are much less accessible to untrained users. Like other aspects of language documentation, the results should be sustainable, which means that they conform to the FAIR principles for data: the data are findable, accessible, interoperable, and reusable. (Of course, this applies only to those languages whose speakers are happy to have descriptions of their languages made available to the whole world without restrictions.)
The dictionaries published in the Dictionaria journal (edited by M. Haspelmath, B. Stiebels and I. Hartmann) have all these advantages, and in addition, they are much more user-friendly than traditional online dictionaries. Below I give an introduction to Dictionaria, and I will sometimes compare the Dictionaria contributions to the online dictionaries of the Webonary series (an SIL International initiative), as well as to the Archi dictionary (created by the Surrey Morphology Group) and the Yurok dictionary (created by linguists at UC Berkeley). The comparisons are merely for illustration, and I do not want to claim that Dictionaria is superior in all respects. But it seems clear that Dictionaria constitutes a great improvement in a number of ways. I will also compare Dictionaria with traditional linear dictionaries, even though it would seem clear that online dictionaries are superior because a dictionary is a database– and databases can be searched by multiple criteria.
Thus, Dictionaria contributions have a view in which they can be easily searched (by typing into the search field at the top of a column). For example, the Nen dictionary (by Nick Evans and colleagues) can be sorted and searched not only by headword, but also by semantic domain – in this view, we see 278 entries showing bird terms (as “BIRD” was selected in the relevant search field), out of a total 5,005 entries.
Of course, all dictionaries give part-of-speech information, but in traditional print dictionaries, this information is given immediately after the headword, and it is not easy to search for all words with a given part of speech. The excerpt here from Terrill Schrock’s linear (open-access) dictionary of Ik illustrates that verbs and nouns appear next to each other.
But some online dictionaries are like print dictionaries, e.g. the Archi dictionary, which does not allow searching by part of speech, but only via the Archi form or the English translation. (The Archi dictionary focuses on presenting inflectional paradigms, something that does not exist in Dictionaria yet.)
The Yurok dictionary allows searching by Yurok word, by English word, by part-of-speech and by semantic domain, but one needs to hit “search” before getting results. If someone already knows that they want to use this dictionary, this is probably not an important limitation, but for potential users who want to get a sense of what kind of information (and how much) the dictionary contains, the tabular presentation of Dictionaria would seem to be much better.
Moreover, in Dictionaria searches, one can specify two conditions simultaneously, e.g. both VERB (part-of-speech) and ‘plant’ (meaning description) in this screenshot from the Iquito dictionary (by Lev Michael and Christine Beier). This yields 34 entries out of 4,128 total entries, which is much more specific than the 586 transitive verbs in Iquito, or the 274 entries which contain ‘plant’ in the meaning description.
Traditional print dictionaries present a “dictionary article” that includes all information on a word in one paragraph, as in the entry qōl ‘surgeonfish’ from the Vurës dictionary (Malau 2021). Such dictionary paragraphs include the part-of-speech and meaning description, plus possibly etymological information, example sentences and a picture.
Such dictionary articles look very nice, so we wanted to have a counterpart of them in Dictionaria, too. Something very similar can be seen in Dictionaria when clicking on the headword. For example, the Teanu word abwa ‘sap of tree’ has the following dictionary entry in Dictionaria (from the dictionary by Alexandre François):
This entry is directly accessible via a URL (https://dictionaria.clld.org/units/teanu-abwa_1), which is an important feature of interoperability. I am not aware of other online dictionaries of small languages that have a unique URL for each entry, though they probably exist.
In order to make dictionary entries less confusing and more readable, some online dictionaries (and even some print dictionaries such as the Vurës dictionary above) use different colours for different text types. For example, the Chepang dictionary of the Webonary series has Chepang in brown, Nepali in blue and English in green, plus grammatical information in orange (only IPA is in black).
But for an online dictionary, it seems simpler to present the different types of information in tabular form, as in Dictionaria. Maybe Webonary chose the different colours and traditional paragraph-based entries in order to make the dictionaries look more familiar. So it is probably true that Dictionaria contributions are unfamiliar in their look, but it should take only some brief testing to see that they are more powerful.
In Dictionaria, we also make some limited use of different colours, for glosses in other languages such as the Bislama glosses in the Daakaka dictionary (by Kilu von Prince). The screenshot below also illustrates another feature of Dictionaria: The examples (may) consist of four lines: the original text, the analyzed text (with hyphens), the gloss line, and the line with the idiomatic translations. Such fully glossed examples are familiar from reference grammars, and there is no reason why dictionaries should not contain fully glossed examples, too.
Example sentences are shown not only in the entries, but also in a table that is accessible by an extra tab “Examples”, and they can be searched just like the entries. For example, the following screenshot shows 9 examples from Daakaka (out of a total of 761) that contain a 1Pl element in the gloss line. The fourth example in this screenshot is shown in the familiar vertically aligned mode (this appears after one clicks on “show” on the right).
Thus, even though Dictionaria contributions require some getting used to, the online version allows everything that all other online dictionaries allow, plus quite a bit more.
An additional unique feature of Dictionaria is that the dictionaries conform to a standard that is getting more and more traction in cross-linguistic studies: the CLDF (Cross-Linguistic Data Format) standard, introduced by Forkel et al. (2018). This makes it possible to store the data independently of the online presentation, and to create additional applications using the same data. The current web application may not exist forever (depending on browser developments and hosting capabilities), but the underlying data exist in a standard format that is also used for other cross-linguistic datasets such as the World atlas of language structures. This is what makes Dictionaria contributions not only user-friendly in the current web application, but also reusable and thus sustainable in the long term.
Forkel, Robert & List, Johann-Mattis & Greenhill, Simon J. & Rzymski, Christoph & Bank, Sebastian & Cysouw, Michael & Hammarström, Harald & Haspelmath, Martin & Kaiping, Gereon A. & Gray, Russell D. 2018. Cross-Linguistic Data Formats, advancing data sharing and re-use in comparative linguistics. Scientific Data 5. 180205. (doi:10.1038/sdata.2018.205)
Malau, Catriona. 2021. A dictionary of Vurës, Vanuatu (Asia-Pacific Linguistics). Canberra, ACT, Australia: Australian National University Press. (https://www.jstor.org/stable/10.2307/j.ctv23hcdvx)
Schrock, Terrill B. 2017. The Ik language: Dictionary and grammar sketch (African Language Grammars and Dictionaries). Berlin: Language Science Press. (http://langsci-press.org/catalog/book/98)