Back to sample AI queries
TypeCompare the overall frequency of all words or phrases in each section of the corpus; for example genres, decades, or dialects
Word/phraseCONJ PRON BE like ,
Corpuscoca
LLM: modelanthropic: claude-3-5
 

1. Data: The construction shows dramatic variation across genres, with highest frequencies in spoken language and TV/Movies, moderate frequencies in magazines and newspapers, and very low frequencies in fiction and academic writing. Temporally, there is a clear and steady increase from 1990 to 2019, with particularly sharp growth after 2005.

2. Possible explanation: The genre distribution strongly suggests this is primarily a feature of informal, conversational language. Its high frequency in spoken contexts and TV/Movies reflects its role in reporting speech and thought in spontaneous discourse. The moderate frequencies in magazines and newspapers likely occur in quoted speech or informal opinion pieces. The very low frequency in academic writing aligns with its informal nature and academic conventions against conversational features. The temporal pattern indicates this is a relatively recent grammatical innovation that has become increasingly conventionalized. Starting as a marginal feature in the early 1990s, it has grown exponentially, particularly in the last decade. The construction's dramatic rise suggests it has become an established way to report speech/thought in informal contexts.

3. Competing constructions: In genres where this construction is infrequent (especially academic and fiction), more traditional quotative verbs likely predominate: - "said" (direct quotation) - "stated" - "reported" - "claimed" - "noted" In fiction, detailed descriptive phrases might be used: - "he replied thoughtfully" - "she responded with hesitation" These alternatives maintain a more formal register appropriate to written genres while providing additional semantic information about the manner of speaking that "be like" doesn't typically convey.