A CORPUS-DRIVEN ANALYSIS OF N-GRAMS IN ENGLISH AND ITALIAN TERMS OF SERVICE
A CORPUS-DRIVEN ANALYSIS OF N-GRAMS IN ENGLISH AND ITALIAN TERMS OF SERVICE

Summary/Abstract: This paper aims at exploring English-Italian equivalences in the N-grams generated from comparable corpora of online terms of service (ToS). To do so, 5-6-word N-grams composed of the most frequent English and Italian nouns, verbs, adjectives and adverbs are analyzed. The English-Italian nouns focused on are the following ones: “service(s)” and “servizi(o)”, “agreement / contract” and “contratto”, and “customer” and “cliente”. The verbs are “be” and “essere”, and “agree” and “accettare”. The adjectives are “such” and “tale”, and the adverbs are “expressly” and “espressamente”. The analysis brings to the fore similarities across the N-grams. For example, the lemmas “terminate” and “recedere” can be considered as equivalents, as well as, to some extent, the verb phrases “agrees to be bound” and “prende atto ed accetta”. In addition, false semantic equivalences appear, such as the N-grams composed of “such” and “tale”, or those made up by “expressly” and “espressamente”. The findings also report the highly repetitiveness of the language of the Italian corpus and, hence, the resulting high frequency of some Italian N-grams. Conversely, the English corpus features more varied lexical patterning with lower frequency rates per N-gram.

• Details
• Contents