Revisiting vocal perception in non-human animals: a review of vowel discrimination, speaker voice recognition, and speaker normalization

Catcalls: exotic cats discriminate the voices of familiar caregivers

Background

The ability to differentiate familiar from unfamiliar humans has been considered a product of domestication or early experience. Few studies have focused on voice recognition in Felidae despite the fact that this family presents the rare opportunity to compare domesticated species to their wild counterparts and to examine the role of human rearing.

Methods

We tested whether non-domesticated Felidae species recognized familiar human voices by exposing them to audio playbacks of familiar and unfamiliar humans. In a pilot study, we presented seven cats of five species with playbacks of voices that varied in familiarity and use of the cats' names. In the main study, we presented 24 cats of 10 species with unfamiliar and then familiar voice playbacks using a habituation-dishabituation paradigm. We anticipated that human rearing and use of the cats' names would result in greater attention to the voices, as measured by the latency, intensity, and duration of responses regardless of subject sex and subfamily.

Results

Cats responded more quickly and with greater intensity (e.g., full versus partial head turn, both ears moved versus one ear twitching) to the most familiar voice in both studies. They also responded for longer durations to the familiar voice compared to the unfamiliar voices in the main study. Use of the cats' name and rearing history did not significantly impact responding. These findings suggest that close human contact rather than domestication is associated with the ability to discriminate between human voices and that less social species may have socio-cognitive abilities akin to those of more gregarious species. With cats of all species being commonly housed in human care, it is important to know that they differentiate familiar from unfamiliar human voices.

Grey wolves (Canis lupus) discriminate between familiar and unfamiliar human voices

The ability to discriminate between different individuals based on identity cues, which is important to support the social behaviour of many animal species, has mostly been investigated in conspecific contexts. A rare example of individual heterospecific discrimination is found in domestic dogs, who are capable of recognising their owners' voices. Here, we test whether grey wolves, the nearest wild relative of dogs, also have the ability to distinguish familiar human voices, which would indicate that dogs' ability is not a consequence of domestication. Using the habituation-dishabituation paradigm, we presented captive wolves with playback recordings of their keepers' and strangers' voices producing either familiar or unfamiliar phrases. The duration of their response was significantly longer when presented with keepers' voices than with strangers' voices, demonstrating that wolves discriminated between familiar and unfamiliar speakers. This suggests that dogs' ability to discriminate between human voices was probably present in their common ancestor and may support the idea that this is a general ability of vertebrates to recognise heterospecific individuals. Our study also provides further evidence for familiar voice discrimination in a wild animal in captivity, indicating that this ability may be widespread across vertebrate species.

The Role of Talking Faces in Infant Language Learning: Mind the Gap between Screen-Based Settings and Real-Life Communicative Interactions

Over the last few decades, developmental (psycho) linguists have demonstrated that perceiving talking faces audio-visually is important for early language acquisition. Using mostly well-controlled and screen-based laboratory approaches, this line of research has shown that paying attention to talking faces is likely to be one of the powerful strategies infants use to learn their native(s) language(s). In this review, we combine evidence from these screen-based studies with another line of research that has studied how infants learn novel words and deploy their visual attention during naturalistic play. In our view, this is an important step toward developing an integrated account of how infants effectively extract audiovisual information from talkers' faces during early language learning. We identify three factors that have been understudied so far, despite the fact that they are likely to have an important impact on how infants deploy their attention (or not) toward talking faces during social interactions: social contingency, speaker characteristics, and task- dependencies. Last, we propose ideas to address these issues in future research, with the aim of reducing the existing knowledge gap between current experimental studies and the many ways infants can and do effectively rely upon the audiovisual information extracted from talking faces in their real-life language environment.

The acoustic bases of human voice identity processing in dogs

Speech carries identity-diagnostic acoustic cues that help individuals recognize each other during vocal–social interactions. In humans, fundamental frequency, formant dispersion and harmonics-to-noise ratio serve as characteristics along which speakers can be reliably separated. The ability to infer a speaker’s identity is also adaptive for members of other species (like companion animals) for whom humans (as owners) are relevant. The acoustic bases of speaker recognition in non-humans are unknown. Here, we tested whether dogs can recognize their owner’s voice and whether they rely on the same acoustic parameters for such recognition as humans use to discriminate speakers. Stimuli were pre-recorded sentences spoken by the owner and control persons, played through loudspeakers placed behind two non-transparent screens (with each screen hiding a person). We investigated the association between acoustic distance of speakers (examined along several dimensions relevant in intraspecific voice identification) and dogs’ behavior. Dogs chose their owner’s voice more often than that of control persons’, suggesting that they can identify it. Choosing success and time spent looking in the direction of the owner’s voice were positively associated, showing that looking time is an index of the ease of choice. Acoustic distance of speakers in mean fundamental frequency and jitter were positively associated with looking time, indicating that the shorter the acoustic distance between speakers with regard to these parameters, the harder the decision. So, dogs use these cues to discriminate their owner’s voice from unfamiliar voices. These findings reveal that dogs use some but probably not all acoustic parameters that humans use to identify speakers. Although dogs can detect fine changes in speech, their perceptual system may not be fully attuned to identity-diagnostic cues in the human voice.

Who is there? Captive western gorillas distinguish human voices based on familiarity and nature of previous interactions

The ability to recognize conspecifics by their acoustic signals is of crucial importance to social animals, especially where visibility is limited, because it allows for discrimination between familiar and unfamiliar individuals and facilitates associations with and the avoidance of particular conspecifics. Animals may also benefit from an ability to recognize and use the information coded into the auditory signals of other species. Companion species such as dogs, cats, and horses are able to discriminate between familiar and unfamiliar human voices; however, whether this ability is widespread across vertebrates is still unknown. Using playback experiments, we tested whether western gorillas living at Zoo Atlanta were able to discriminate between the voices of subgroups of people: i.e., unfamiliar individuals, familiar individuals with whom the gorillas had positive interactions, and familiar individuals with whom they had negative interactions. Gorillas responded significantly more often (longer gazing duration, higher gazing frequency, shorter latency, and larger number of distress behaviors) to the voices of unfamiliar and familiar-negative individuals than to those of familiar-positive individuals, indicating that they recognized the voices of subgroup of people based on familiarity and possibly the nature of the relationship with them. Future studies should determine whether this is also the case in the wild, where interspecific associations with humans are less intense than they are in captive settings.

Cross-Dialectal Novel Word Learning and Borrowing

The objective of this paper was to study the cognitive processes underlying cross-dialectal novel word borrowing and loanword establishment in a Standard-Chinese-to-Shanghainese (SC-SH) auditory lexical learning and borrowing experiment. To investigate these underlying cognitive processes, SC-SH bi-dialectals were compared with SC monolectals as well as bi-dialectals of SC and other Chinese dialects (OD) to investigate the influence of short-term and long-term linguistic experience. Both comprehension and production borrowings were tested. This study found that early and proficient bi-dialectism, even if it is not directly related to the recipient dialect of lexical borrowing, has a protective effect on the ability of cross-dialectal lexical borrowing in early adulthood. Bi-dialectals tend to add separate lexical representations for incidentally encountered dialectal variants, while monolectals tend to assimilate dialectal variants to standard forms. Bi-dialectals, but not monolectals, use etymologically related morphemes between the source and recipient dialects to create nonce-borrowing compounds. Dialectal variability facilitates lexical borrowing via enriching instead of increasing the short-term lexical experience of learners. The long-term bi-dialectal experience of individuals, as well as their short-term exposure to each specific loanword, may collectively shape the route of lexical evolution of co-evolving linguistic varieties.

Who is there? Captive western gorillas distinguish human voices based on familiarity and nature of previous interactions

The ability to recognize conspecifics by their acoustic signals is of crucial importance to social animals, especially where visibility is limited, because it allows for discrimination between familiar and unfamiliar individuals and facilitates associations with and the avoidance of particular conspecifics. Animals may also benefit from an ability to recognize and use the information coded into the auditory signals of other species. Companion species such as dogs, cats, and horses are able to discriminate between familiar and unfamiliar human voices; however, whether this ability is widespread across vertebrates is still unknown. Using playback experiments, we tested whether western gorillas living at Zoo Atlanta were able to discriminate between the voices of subgroups of people: i.e., unfamiliar individuals, familiar individuals with whom the gorillas had positive interactions, and familiar individuals with whom they had negative interactions. Gorillas responded significantly more often (longer gazing duration, higher gazing frequency, shorter latency, and larger number of distress behaviors) to the voices of unfamiliar and familiar-negative individuals than to those of familiar-positive individuals, indicating that they recognized the voices of subgroup of people based on familiarity and possibly the nature of the relationship with them. Future studies should determine whether this is also the case in the wild, where interspecific associations with humans are less intense than they are in captive settings.

Repetition enhancement to voice identities in the dog brain

In the human speech signal, cues of speech sounds and voice identities are conflated, but they are processed separately in the human brain. The processing of speech sounds and voice identities is typically performed by non-primary auditory regions in humans and non-human primates. Additionally, these processes exhibit functional asymmetry in humans, indicating the involvement of distinct mechanisms. Behavioural studies indicate analogue side biases in dogs, but neural evidence for this functional dissociation is missing. In two experiments, using an fMRI adaptation paradigm, we presented awake dogs with natural human speech that either varied in segmental (change in speech sound) or suprasegmental (change in voice identity) content. In auditory regions, we found a repetition enhancement effect for voice identity processing in a secondary auditory region - the caudal ectosylvian gyrus. The same region did not show repetition effects for speech sounds, nor did the primary auditory cortex exhibit sensitivity to changes either in the segmental or in the suprasegmental content. Furthermore, we did not find evidence for functional asymmetry neither in the processing of speech sounds or voice identities. Our results in dogs corroborate former human and non-human primate evidence on the role of secondary auditory regions in the processing of suprasegmental cues, suggesting similar neural sensitivity to the identity of the vocalizer across the mammalian order.

Dogs perceive and spontaneously normalize formant-related speaker and vowel differences in human speech sounds

Domesticated animals have been shown to recognize basic phonemic information from human speech sounds and to recognize familiar speakers from their voices. However, whether animals can spontaneously identify words across unfamiliar speakers (speaker normalization) or spontaneously discriminate between unfamiliar speakers across words remains to be investigated. Here, we assessed these abilities in domestic dogs using the habituation-dishabituation paradigm. We found that while dogs habituated to the presentation of a series of different short words from the same unfamiliar speaker, they significantly dishabituated to the presentation of a novel word from a new speaker of the same gender. This suggests that dogs spontaneously categorized the initial speaker across different words. Conversely, dogs who habituated to the same short word produced by different speakers of the same gender significantly dishabituated to a novel word, suggesting that they had spontaneously categorized the word across different speakers. Our results indicate that the ability to spontaneously recognize both the same phonemes across different speakers, and cues to identity across speech utterances from unfamiliar speakers, is present in domestic dogs and thus not a uniquely human trait.

Mechanisms underlying speech sound discrimination and categorization in humans and zebra finches

Speech sound categorization in birds seems in many ways comparable to that by humans, but it is unclear what mechanisms underlie such categorization. To examine this, we trained zebra finches and humans to discriminate two pairs of edited speech sounds that varied either along one dimension (vowel or speaker sex) or along two dimensions (vowel and speaker sex). Sounds could be memorized individually or categorized based on one dimension or by integrating or combining both dimensions. Once training was completed, we tested generalization to new speech sounds that were either more extreme, more ambiguous (i.e., close to the category boundary), or within-category intermediate between the trained sounds. Both humans and zebra finches learned the one-dimensional stimulus-response mappings faster than the two-dimensional mappings. Humans performed higher on the trained, extreme and within-category intermediate test-sounds than on the ambiguous ones. Some individual birds also did so, but most performed higher on the trained exemplars than on the extreme, within-category intermediate and ambiguous test-sounds. These results suggest that humans rely on rule learning to form categories and show poor performance when they cannot apply a rule. Birds rely mostly on exemplar-based memory with weak evidence for rule learning.

Formant-frequency discrimination of synthesized vowels in budgerigars (Melopsittacus undulatus) and humans

Vowels are complex sounds with four to five spectral peaks known as formants. The frequencies of the two lowest formants, F1and F2, are sufficient for vowel discrimination. Behavioral studies show that many birds and mammals can discriminate vowels. However, few studies have quantified thresholds for formant-frequency discrimination. The present study examined formant-frequency discrimination in budgerigars (Melopsittacus undulatus) and humans using stimuli with one or two formants and a constant fundamental frequency of 200 Hz. Stimuli had spectral envelopes similar to natural speech and were presented with random level variation. Thresholds were estimated for frequency discrimination of F1, F2, and simultaneous F1 and F2 changes. The same two-down, one-up tracking procedure and single-interval, two-alternative task were used for both species. Formant-frequency discrimination thresholds were as sensitive in budgerigars as in humans and followed the same patterns across all conditions. Thresholds expressed as percent frequency difference were higher for F1 than for F2, and were unchanged between stimuli with one or two formants. Thresholds for simultaneous F1 and F2 changes indicated that discrimination was based on combined information from both formant regions. Results were consistent with previous human studies and show that budgerigars provide an exceptionally sensitive animal model of vowel feature discrimination.

Selective auditory grouping by zebra finches: testing the iambic-trochaic law

Humans have a strong tendency to spontaneously group visual or auditory stimuli together in larger patterns. One of these perceptual grouping biases is formulated as the iambic/trochaic law, where humans group successive tones alternating in pitch and intensity as trochees (high-low and loud-soft) and alternating in duration as iambs (short-long). The grouping of alternations in pitch and intensity into trochees is a human universal and is also present in one non-human animal species, rats. The perceptual grouping of sounds alternating in duration seems to be affected by native language in humans and has so far not been found among animals. In the current study, we explore to which extent these perceptual biases are present in a songbird, the zebra finch. Zebra finches were trained to discriminate between short strings of pure tones organized as iambs and as trochees. One group received tones that alternated in pitch, a second group heard tones alternating in duration, and for a third group, tones alternated in intensity. Those zebra finches that showed sustained correct discrimination were next tested with longer, ambiguous strings of alternating sounds. The zebra finches in the pitch condition categorized ambiguous strings of alternating tones as trochees, similar to humans. However, most of the zebra finches in the duration and intensity condition did not learn to discriminate between training stimuli organized as iambs and trochees. This study shows that the perceptual bias to group tones alternating in pitch as trochees is not specific to humans and rats, but may be more widespread among animals.

Cross-Modal Correspondences in Non-human Mammal Communication

For both humans and other animals, the ability to combine information obtained through different senses is fundamental to the perception of the environment. It is well established that humans form systematic cross-modal correspondences between stimulus features that can facilitate the accurate combination of sensory percepts. However, the evolutionary origins of the perceptual and cognitive mechanisms involved in these cross-modal associations remain surprisingly underexplored. In this review we outline recent comparative studies investigating how non-human mammals naturally combine information encoded in different sensory modalities during communication. The results of these behavioural studies demonstrate that various mammalian species are able to combine signals from different sensory channels when they are perceived to share the same basic features, either because they can be redundantly sensed and/or because they are processed in the same way. Moreover, evidence that a wide range of mammals form complex cognitive representations about signallers, both within and across species, suggests that animals also learn to associate different sensory features which regularly co-occur. Further research is now necessary to determine how multisensory representations are formed in individual animals, including the relative importance of low level feature-related correspondences. Such investigations will generate important insights into how animals perceive and categorise their environment, as well as provide an essential basis for understanding the evolution of multisensory perception in humans.

Emotional and Interactional Prosody across Animal Communication Systems: A Comparative Approach to the Emergence of Language

Across a wide range of animal taxa, prosodic modulation of the voice can express emotional information and is used to coordinate vocal interactions between multiple individuals. Within a comparative approach to animal communication systems, I hypothesize that the ability for emotional and interactional prosody (EIP) paved the way for the evolution of linguistic prosody - and perhaps also of music, continuing to play a vital role in the acquisition of language. In support of this hypothesis, I review three research fields: (i) empirical studies on the adaptive value of EIP in non-human primates, mammals, songbirds, anurans, and insects; (ii) the beneficial effects of EIP in scaffolding language learning and social development in human infants; (iii) the cognitive relationship between linguistic prosody and the ability for music, which has often been identified as the evolutionary precursor of language.

Speaker and Accent Variation Are Handled Differently: Evidence in Native and Non-Native Listeners

Listeners are able to cope with between-speaker variability in speech that stems from anatomical sources (i.e. individual and sex differences in vocal tract size) and sociolinguistic sources (i.e. accents). We hypothesized that listeners adapt to these two types of variation differently because prior work indicates that adapting to speaker/sex variability may occur pre-lexically while adapting to accent variability may require learning from attention to explicit cues (i.e. feedback). In Experiment 1, we tested our hypothesis by training native Dutch listeners and Australian-English (AusE) listeners without any experience with Dutch or Flemish to discriminate between the Dutch vowels /I/ and /ε/ from a single speaker. We then tested their ability to classify /I/ and /ε/ vowels of a novel Dutch speaker (i.e. speaker or sex change only), or vowels of a novel Flemish speaker (i.e. speaker or sex change plus accent change). We found that both Dutch and AusE listeners could successfully categorize vowels if the change involved a speaker/sex change, but not if the change involved an accent change. When AusE listeners were given feedback on their categorization responses to the novel speaker in Experiment 2, they were able to successfully categorize vowels involving an accent change. These results suggest that adapting to accents may be a two-step process, whereby the first step involves adapting to speaker differences at a pre-lexical level, and the second step involves adapting to accent differences at a contextual level, where listeners have access to word meaning or are given feedback that allows them to appropriately adjust their perceptual category boundaries.

Phonological perception by birds: budgerigars can perceive lexical stress

Metrical phonology is the perceptual "strength" in language of some syllables relative to others. The ability to perceive lexical stress is important, as it can help a listener segment speech and distinguish the meaning of words and sentences. Despite this importance, there has been little comparative work on the perception of lexical stress across species. We used a go/no-go operant paradigm to train human participants and budgerigars (Melopsittacus undulatus) to distinguish trochaic (stress-initial) from iambic (stress-final) two-syllable nonsense words. Once participants learned the task, we presented both novel nonsense words, and familiar nonsense words that had certain cues removed (e.g., pitch, duration, loudness, or vowel quality) to determine which cues were most important in stress perception. Members of both species learned the task and were then able to generalize to novel exemplars, showing categorical learning rather than rote memorization. Tests using reduced stimuli showed that humans could identify stress patterns with amplitude and pitch alone, but not with only duration or vowel quality. Budgerigars required more than one cue to be present and had trouble if vowel quality or amplitude were missing as cues. The results suggest that stress patterns in human speech can be decoded by other species. Further comparative stress-perception research with more species could help to determine what species characteristics predict this ability. In addition, tests with a variety of stimuli could help to determine how much this ability depends on general pattern learning processes versus vocalization-specific cues.

A general auditory bias for handling speaker variability in speech? Evidence in humans and songbirds

Different speakers produce the same speech sound differently, yet listeners are still able to reliably identify the speech sound. How listeners can adjust their perception to compensate for speaker differences in speech, and whether these compensatory processes are unique only to humans, is still not fully understood. In this study we compare the ability of humans and zebra finches to categorize vowels despite speaker variation in speech in order to test the hypothesis that accommodating speaker and gender differences in isolated vowels can be achieved without prior experience with speaker-related variability. Using a behavioral Go/No-go task and identical stimuli, we compared Australian English adults’ (naïve to Dutch) and zebra finches’ (naïve to human speech) ability to categorize / I/ and /ε/ vowels of an novel Dutch speaker after learning to discriminate those vowels from only one other speaker. Experiments 1 and 2 presented vowels of two speakers interspersed or blocked, respectively. Results demonstrate that categorization of vowels is possible without prior exposure to speaker-related variability in speech for zebra finches, and in non-native vowel categories for humans. Therefore, this study is the first to provide evidence for what might be a species-shared auditory bias that may supersede speaker-related information during vowel categorization. It additionally provides behavioral evidence contradicting a prior hypothesis that accommodation of speaker differences is achieved via the use of formant ratios. Therefore, investigations of alternative accounts of vowel normalization that incorporate the possibility of an auditory bias for disregarding inter-speaker variability are warranted.

Zebra finches are able to learn affixation-like patterns

Adding an affix to transform a word is common across the world languages, with the edges of words more likely to carry out such a function. However, detecting affixation patterns is also observed in learning tasks outside the domain of language, suggesting that the underlying mechanism from which affixation patterns have arisen may not be language or even human specific. We addressed whether a songbird, the zebra finch, is able to discriminate between, and generalize, affixation-like patterns. Zebra finches were trained and tested in a Go/Nogo paradigm to discriminate artificial song element sequences resembling prefixed and suffixed ‘words.’ The ‘stems’ of the ‘words,’ consisted of different combinations of a triplet of song elements, to which a fourth element was added as either a ‘prefix’ or a ‘suffix.’ After training, the birds were tested with novel stems, consisting of either rearranged familiar element types or novel element types. The birds were able to generalize the affixation patterns to novel stems with both familiar and novel element types. Hence, the discrimination resulting from the training was not based on memorization of individual stimuli, but on a shared property among Go or Nogo stimuli, i.e., affixation patterns. Remarkably, birds trained with suffixation as Go pattern showed clear evidence of using both prefix and suffix, while those trained with the prefix as the Go stimulus used primarily the prefix. This finding illustrates that an asymmetry in attending to different affixations is not restricted to human languages.