Intonation contours reliably convey meaning intended by the talker (Ladd, 2008). Acoustic realizations of intonation contours are notoriously variable, however (Arvaniti, 2019). Contours treated by speakers and listeners as instances of the “same” tune can differ substantially from each other, both for systematic factors such as talkers (e.g., age, gender, dialect) and speaking conditions, as well as for random factors (Clopper & Smiljanic, 2011). Currently, little is known about how listeners can resolve the uncertainty resulting from the variable acoustic input. The current study examines /talker variability/ associated with two utterance types – rising (e.g., /It’s raining?/) vs. falling (e.g., /It’s raining/) declaratives. These are well established intonational phenomena (Gunlogson, 2008; Jeong & Potts, 2016), often considered to be perceptually separable from one another. We first ask how and how much talkers differ from one another and how listeners may be able to overcome the talker-variability. We then adopt Ideal Observer analyses (Kleinschmidt, 2019) to test whether the structure of the variability can /in principle/ afford listeners the means to navigate the variability in intonation processing.