Aging alters various aspects of language output. However, health care professionals lack knowledge about language skills in the vastly growing very old (VO) population. Narratives provide a lucrative method for examining language in ageing as they are sensitive to neurological changes.

We study the normative lexical diversity and pauses in healthy 80–100-year-old speakers’ narratives and investigate whether they are predicted by physio-anatomical, socio-cognitive or linguistic background variables.

This cross-sectional study comprises fifty VO Finnish speakers, who produced a narrative from a wordless cartoon strip. The lexical diversity of the narratives was analysed by number of different words (NDW), type-token-ratio (TTR), words by grammatical class and noun-verb ratio. Pausal variables consisted of pause time ratio, pause frequency, mean pause duration and mean syllables between pauses. Background variables included age, dentition (natural teeth vs. dentures), educational level and linguistic skills by the Western Aphasia Battery Aphasia quotient (WAB-AQ), semantic and phonemic fluency and the Boston Naming Test (BNT). The association between lexical diversity, pausing and the background variables was studied by multiple regression analysis.

Both lexical diversity and pausing showed notable individual variation. Higher NDW was linked to natural teeth among lexical diversity variables. NDW was also positively predicted by the highest educational level and WAB-AQ but negatively associated with BNT scores. Within the pause variables, pause time ratio was predicted by age as the oldest speakers produced the lowest proportions of pauses.

This study offers normative data on multiple variables of the VO speakers’ language use. It also provides insight into the interconnections between ageing, language and personal factors. As natural teeth and high education were associated with diverse vocabulary, we recommend that clinicians assess the participants’ dental status and educational history. The contrasting effects of the WAB-AQ and BNT on lexical diversity underscore the need to acknowledge that lexical abilities may appear differently in formal tests and in narration. The finding that the oldest of the speakers used less time pausing challenges the assumption about pausing consistently increasing with age. This study adds to the missing knowledge on ageing and language. It provides health care professionals more expertise in identifying typical age-related variety. Nevertheless, to enable narrative evaluation in routine clinical practice more international research on typical old and adult speakers is required as well as developments in automated analysis of spoken language.