Effects of sub-optimal housing on inactivity vary across species and experiments, probably because inactivity is heterogeneous, reflecting both positive states (e.g. relaxation) and negative ones (e.g. fear). We therefore aimed to identify specific subtypes of inactivity that could indicate poor welfare in mink, by comparing their behaviour in enriched and non-enriched conditions (the former having been previously demonstrated to be highly preferred by mink and to enhance their welfare). We assessed this in three groups of subjects, as well as after housing conditions were reversed for the last group. During live scans, inactive animals were scored for posture, location, and whether awake or apparently asleep. Data on temperament and physiological stress indicators were also collected for one group; these confirmed that non-enriched housing increased faecal cortisol metabolites (FCM; P = 0.040). Non-enriched housing also increased locomotor stereotypy in females (sex × housing: P = 0.004). Inactivity in the nest-box (vs. in the open cage) was higher among females in non-enriched housing (housing × sex: P < 0.001), and increased by 20% of observations after enrichment removal (P = 0.018) for both sexes. Furthermore, males with fearful temperaments spent the most time inactive in the nest-box (sex × temperament: P = 0.054), while females whose FCM decreased most when given enrichment also showed the largest decreases in this behaviour (sex × FCM change: P = 0.019). Together, this suggests that inactivity in the nest-box may reflect anxiety-induced hiding. Lying awake (i.e. prone with eyes open) was also higher in non-enriched housing (3.1% of observations vs. 1.7%; P = 0.002); furthermore, this subtype of inactivity increased after enrichment removal (by 1.0% of observations; P = 0.021), and decreased when non-enriched mink were given enrichment (by 2.4% of observations; P = 0.004). This behaviour did not co-vary with fearfulness, however, nor with FCM (both P > 0.05). This suggests that lying awake is not fear-related (e.g. not reflecting enhanced vigilance) but instead reflects some other negative state. Effects on inactivity subtypes as defined by posture were less consistent. For example, time spent lying belly down tended to decrease in mink moved from non-enriched to enriched cages (P = 0.054), but enriched mink spent significantly less time belly down (in one of the three groups; P = 0.002). Overall, two subtypes of inactivity, lying in the nest-box and lying awake seem likely to be valid indicators of housing-induced poor welfare in this species, being consistently increased by non-enriched cages. Lying in the nest-box may indicate fear or anxiety, and lying awake, a boredom-like state.