**Equipment**: generic cheap clip on mic connected to smartphone. Recording app: ‘Voice Record’ by Dayana Networks ltd (for Melanie’s lab) **Target objects**: 5 familiar and 5 unfamiliar noise-free objects with variation in phonemes, that aren’t usually “y” words and are 1-syllable: *Familiar set would be*: ball, shoe, block, cup, train *Unfamiliar set would be*: sieve, globe, whisk, flag, yeast See folder named, 'Objects Soderstrom' in Files. **Location**: Child-friendly room. Quiet, but not sound-proofed. For the ID sample, mom and baby with research assistant, who will give out the objects. For the AD sample, just mom and an experimenter (need a second RA to watch the baby for five minutes). **Participants**: Mothers (English monolinguals) with their 4-7 month old infant. **Instructions/Protocol**: Mothers are told that we want to record them interacting with their baby and with an adult. We tell them ahead of time that we are going to use these recordings in a perceptual study with infants, and that the recordings will be made available to other researchers. We give them a bag with the objects and have a research assistant pull one object out of the bag at a time and hand it to the mother to talk about the objects with their baby/the adult until they run out of things to say and then pull out the next object. Order of baby first or adult first is counterbalanced across participants. Order of objects is random based on what gets pulled out of the bag. Try to get at least 5 minutes worth of speech, but record until they run out of things to say.