PLC Analysis

synNetAn analysis of the Pattern of Linguistic Complexity – or PLC analysis – provides of important information about how the speaker combines words.
Typically, humans perfom in a way that a general common pattern can be extracted from their performance. However, when there is a disturbance that affect their linguistic capability, the pattern is also affected. This is particularly clear in the case of aphasias or genetic-based syndromes (Down syndrome, Williams syndrome, etc).

In order to do a PLC analysis we use the Netlang software, a tool developed in 2014-2015. This is a program with which it is possible to make a linguistic analysis and then send it it to a network program.

Netlang is based on Dependency Grammar. Information about the program has been published here:

(2017) Barceló-Coblijn L, Serna Salazar D, Isaza G, Castillo Ossa LF, Bedia MG (2017) Netlang: A software for the linguistic analysis of corpora by means of complex networks. PLoS ONE 12(8): e0181341.

journal.pone.0181341.g002

(A) Screenshot of Netlang software as it is. (A 1) subwindow showing the analyzed text; (A 2) subwindow showing the original text; (A 3) subwindow showing the gathering of the three columns of data; (A 4) window for a first representation of the network; (A 5) internal dependency grammar of Netlang; (A 6) in this bare appear all analyzed sentences and the user can select them in order to edit; (B) emergent window from the Bookannotations menu that allows to set up the labels for the dependency relations; (C) emergent window from the Bookannotations menu for the customization of the labels of word categories; (D) emergent Netlang window where syntactic relations have been established between words, by means of arrows which include a syntactic label (e.g., «modifier» or «subject»). Lexical categories have also been identified under the words (e.g., «preposition», «article»). (E) emergent window where the user can check the text to be analyzed and modify it if necessary. 

And this is the final result (in this case, the network has been visualized with Cytoscape).

xarxa1

 

It is also possible to analyze bilingual corpora. In doing this one will get a bilingual network. One possibility, for example, is editing lexical labels and classifying all words according to the language they belong to. Here you are an example of a bilingual network of a child acquiring English (in red) and Spanish (in grey; in blue, proper names) [pdf]:

870823ej SOLID
Anuncis