The human ability to detect regularities in sound sequences is a fundamental substrate of our language faculty. However, is this an ability exclusive to human language processing, or have we usurped a more general learning mechanism for this purpose, one shared with other species? The current study is an attempt to replicate and extend Hauser, Weiss, and Marcus’s (2002) retracted study (2010) of artificial grammar learning in tamarins to determine if tamarins can detect an underlying grammatical structure in a pattern of sounds. Human language consonant–vowel (CV) combinations from Hauser et al.’s original study, newly created tone sequences, and newly created monkey vocalizations made into sequences were used to familiarize tamarins to an AAB or ABB pattern. Tests of novel sounds in each condition were presented that either were consistent with the familiarized pattern or were different from it. Longer looking times toward the sound source (an audio speaker with a specific location in the auditory field) indicated recognition of novelty. Tamarins looked toward the speaker significantly longer with inconsistent human language CV sequences and with inconsistent tone sequences but not when an inconsistent monkey vocalization was presented. Moreover, tamarins showed differential rates of habituation to the different types of sound patterns, with more robust habituation to CV sequences and tone sequences than to monkey call sequences. The implications of these findings for the generality of learning mechanisms for linguistic and nonlinguistic input across species and the importance of testing across various stimuli are discussed.