backshifted

Backshifted is an English adjective used in linguistics and time series analysis to describe something that has been moved backward in time or sequence.

In linguistics, backshifted refers to the processing that occurs when converting direct speech into reported or indirect speech. Tenses and certain modal verbs are often shifted back in time. For example, "She says, 'I am tired'" becomes "She says she is tired" in some contexts, but in classic backshifting, "I am tired" becomes "she said she was tired" and "I will go" becomes "she said she would go." Modal verbs commonly change as well: can → could, may → might, will → would. Time expressions typically move backward, so “tomorrow” becomes “the next day.” There is nuance: if the truth of a statement is still the case, some contexts retain the present tense instead of backshifting.

In time series and econometrics, backshifted describes data that have been shifted backward by one or more time periods. This is done using the backshift operator, often denoted B. If X_t is a series, the backshifted value is X_{t-1}; more generally, B^k X_t = X_{t-k}. Backshifted series are used to express lagged relationships in models such as ARIMA and to analyze delayed effects.

Across domains, the core idea is moving data or language frames toward the past. The term emphasizes a temporal realignment that affects tense, meaning, or data indexing.