Dialectic signatures in animal acoustic signals are key in the identification of and association with group members. Complex vocal sequences may also convey information about behavioral state, and may thus vary according to social environment. Some bird species, such as psittaciforms, learn and modify their complex acoustic signals throughout their lives. However, the structure and function of vocal sequences in open-ended vocal learners remains understudied. Here, we examined vocal sequence variation in the warble song of budgerigars, and how these change upon contact between social groups. Budgerigars are open-ended vocal learners which exhibit fission–fusion flock dynamics in the wild. We found that two captive colonies of budgerigars exhibited colony-specific differences in the syntactic structure of their vocal sequences. Individuals from the two colonies differed in the propensity to repeat certain note types, forming repetitive motifs which served as higher-order signatures of colony identity. When the two groups were brought into contact, their vocal sequences converged, and these colony-specific repetitive patterns disappeared, with males from both erstwhile colonies now producing similar sequences with similar syntactic structure. We present data suggesting that the higher-order temporal arrangement of notes/vocal units is modified throughout life by social learning as groups of birds continually associate and dissociate. Our study sheds light on the importance of examining signal structure at multiple levels of organization, and the potential for psittaciform birds as model systems to examine the influence of learning and social environment on acoustic signals.

Animal sensory signals are often complex and elaborate, conveying diverse information such as behavioral context, mate quality, presence of a predator, territoriality and group membership (Bradbury and Vehrencamp, 2011). Diverse animal taxa employ sound in the form of acoustic signals to communicate information over both shorter and longer distances. Complex signals support communication of contextual information, by changes in the content and structure of signals (Bradbury and Vehrencamp, 2011; Engesser et al., 2019; Hebets and Papaj, 2005; Suzuki et al., 2018). For example, members of a species may change the structure of a single element in a signal, or the temporal structure (syntax) of an acoustic signal sequence to convey information (Bhat et al., 2022; Bohn et al., 2009; Engesser et al., 2019; Leroux et al., 2021; Suzuki et al., 2018; Zuberbühler, 2018). The larger the repertoire of elements, the greater the combinatorial diversity of sequences, and the more complex the information content of signals (Backhouse et al., 2022; Balsby et al., 2017; Clay and Zuberbühler, 2011; Dalziell and Welbergen, 2016; Leroux et al., 2021; Mitoyen et al., 2019; Scholes, 2008). In addition to behavioral state, group-living animals may use complex signals to communicate group or individual identity (Mammen and Nowicki, 1981; McComb et al., 2000; McCowan and Hooper, 2002). The dialects of passerine songbirds represent one such form of variation, with individuals exhibiting discrimination between familiar and unfamiliar dialects (Briefer et al., 2013; Farabaugh et al., 1988; Henry et al., 2015; Marler and Slabbekoorn, 2004; Marler and Tamura, 1962; Searcy et al., 1981). Oscine passerines, in addition to certain other bird groups such as parrots, learn their complex vocal signals. Therefore, both the individual vocal units (notes or syllables) and the syntactic structure of the song are transmitted by diverse learning mechanisms (Hultsch and Todt, 1989; Lipkind et al., 2017). The structure of songs carries signatures of population, species and individual identity (Allen et al., 2018; Deecke et al., 2010; Marler and Tamura, 1962; Peters et al., 1980; Rendell and Whitehead, 2003).

Unlike most well-studied passerine systems, members of the parrot order Psittaciformes are open-ended vocal learners, modifying their vocal repertoires throughout their lives. Various parrot species are highly social, and can rapidly modify their vocalizations, which is proposed to mediate fission–fusion dynamics in flocking species (Berg et al., 2012; Dahlin et al., 2014; Hile et al., 2000, 2005). For example, when smaller flocks aggregate into larger ones, or individuals move between flocks, the acoustic structure of individual flight call notes changes to match that of the new group (Dahlin et al., 2014; Farabaugh et al., 1994; Hile and Striedter, 2000; Hile et al., 2000, 2005; Wanker et al., 2005). This modification or vocal convergence is hypothesized to mediate affiliative and aggressive interactions between different flock members, or in pair bonding and pair-bond maintenance (Dahlin et al., 2014; Hile et al., 2000, 2005; Mammen and Nowicki, 1981; Tyack, 2008). Similar convergence is observed in primates, and is thought to serve similar social functions (Candiotti et al., 2012; Elowson and Snowdon, 1994; Mitani and Gros-Luis, 1998). However, the individual call notes of parrots are often emitted as part of highly complex vocal sequences (Balsby et al., 2017; Farabaugh et al., 1992; Zdenek et al., 2015). Although the studies discussed above have identified convergence in the acoustic structure of individual call notes due to pair bonding or mixing of flocks, the vocal sequences of parrots remain sparsely studied. In particular, there is no information on how the syntactic structure of vocal sequences is modified by social context, and whether group identity is conveyed by higher-order information within sequences as compared with individual notes. Further, there is almost no information on how the structure of vocal sequences changes when different groups come into contact.

The budgerigar (Melopsittacus undulatus) is a social psittaciform species native to Australia, which has been extensively domesticated for over a century. Male budgerigars typically emit a complex, lengthy sequence of pre-copulatory vocalizations, called a warble song (Fig. 1A), which is sung both to other males and to females (Brockway, 1964; Farabaugh et al., 1992). Although individuals from different colonies differ in the acoustic structure of individual call notes, which converge when individuals are introduced into a new social group (Dahlin et al., 2014; Farabaugh et al., 1994; Hile and Striedter, 2000), the sequence of the warble song has received very little study in these contexts. Here, we examined the syntactic structure of the warble song in laboratory colonies of budgerigars to understand whether the complexity and syntax of the warble varies according to social context and group identity. Specifically, we tested whether warble syntax differs between male–male and male–female contexts, and also whether syntactic structure encodes the identity of the colony from which the male originates. Next, we examined whether the warble song exhibits syntactic convergence when groups of birds are brought into contact. We predicted that warble sequences could exhibit either context- or colony-specific variation. Further, in a situation where warble structure encodes colony identity, we predicted that the syntactic structure of warbles should exhibit greater similarity after contact than before. Our study thus examined the presence of higher-order dialectic variation and vocal convergence in psittaciform birds, which has implications for our understanding of complex communication in this sophisticated signaling system. In addition, open-ended vocal learners such as budgerigars provide valuable insight into how other complex acoustic signals (including human language) exhibit higher-order change due to social learning.

Fig. 1.

The warble song of budgerigars. (A) Spectrograms representing a total of 20.53 s of a warble song produced by a male budgerigar, with note types annotated below the spectrogram. (B) Examples of the seven different note types in budgerigar warble song. ‘a’, alarm call-like elements; ‘b’, contact call-like elements; ‘c’, long harmonic sounds with duration more than 100 ms; ‘d’, short harmonic sounds with duration less than 100 ms; ‘e’, ‘noisy’ calls – broadband sounds; ‘f’, short-duration broadband ‘clicks’; ‘m’, ‘soft calls’ – low-amplitude sounds. See Materials and Methods for further details.

Fig. 1.

The warble song of budgerigars. (A) Spectrograms representing a total of 20.53 s of a warble song produced by a male budgerigar, with note types annotated below the spectrogram. (B) Examples of the seven different note types in budgerigar warble song. ‘a’, alarm call-like elements; ‘b’, contact call-like elements; ‘c’, long harmonic sounds with duration more than 100 ms; ‘d’, short harmonic sounds with duration less than 100 ms; ‘e’, ‘noisy’ calls – broadband sounds; ‘f’, short-duration broadband ‘clicks’; ‘m’, ‘soft calls’ – low-amplitude sounds. See Materials and Methods for further details.

Study animals

We carried out our experiments at the Indian Institute of Science Education and Research (IISER) campuses in Pune and Bhopal. All procedures were approved by the Institutional Animal Ethics Committees of both institutes (protocol numbers: IISER_Pune/IAEC/2020_01_01 and IISERB/2022/001) in accordance with the guidelines laid out by the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA, New Delhi). All our experiments were carried out on captive budgerigars, Melopsittacus undulatus (Shaw 1805), procured from local pet stores, which were sexed by differences in cere color. In our study, colony 1 consisted of two males and four females procured in March 2020, and colony 2 consisted of three males, one female and one bird whose sex could not be conclusively determined by cere color (all pairs containing this bird were omitted from comparisons of male–male and male–female contexts, but were included in colony-level comparisons), procured from a different supplier in August 2021. Thus, the two colonies originated from different sources. We housed the birds in separate rooms within the same lab at IISER Pune for the first part of the study, thus ensuring no visual or acoustic contact between them. We provided the birds with ad libitum water and commercially available bird seed during all experiments, and housed them under a 12 h:12 h light:dark cycle using fluorescent lamps. The rooms were maintained at 25°C using an air conditioner, and birds were provided with boiled eggs or coriander leaves every few days to supplement their diet. The colony cages contained perches at different heights, as well as facilities for the birds to climb around the cage. In late December 2021, the bird colony moved to IISER Bhopal, and the two colonies were combined into the same cages and housed together during and after this move. This enabled us to mimic the fission–fusion dynamics of natural budgerigar flocks and examine its effect on vocal sequences and syntax.

Experimental procedure and recording

Colony 1 contained one possible male–male pair and eight possible male–female pairs (based on the number of possible combinations of individuals within a colony), whereas colony 2 possessed three possible male–male and three possible male–female pairs (and also three pairs including the males with the one bird of indeterminate sex). In order to examine whether social context shaped the sequences of notes emitted during warble song, we placed each of these within-colony pairs inside a soundproof acoustic enclosure of dimensions 75 cm×75 cm×75 cm (Newtech Engineering Systems, Bengaluru, Karnataka, India). This enclosure was acoustically isolated from the external environment. Birds remained in the experimental enclosure for 24 h, to habituate them to the experimental paradigm. We then recorded budgerigar vocalizations for the next 24 h at a sampling rate of 44.1 kHz, using Audiomoth recorders (Hill et al., 2019). The recordings were carried out in September and October 2021. As described above, we recorded warble songs from a total of nine pairs from colony 1 and nine pairs from colony 2. For the second part of the study, our goal was to examine whether syntax had changed after the two colonies came into contact with each other. Thus, we re-recorded all pairs in June 2022 and January 2023, after the colonies were brought into contact. We had two fewer male–female pairs in colony 1 for the dataset collected after contact owing to the loss of female birds from natural causes in the intervening period and during data collection. We therefore analyzed additional warbles of males from the erstwhile colony 1, to ensure that the two datasets had an equal number of warbles per male. Additionally, one male from the erstwhile colony 1 died of natural causes before we re-recorded male–male pairs in January 2023. Therefore, we restricted comparisons based on context after contact to pairs in colony 2, for which we had multiple possible pair combinations. Our focus broadly was on colony-level warble patterns, but wherever we compared individual birds with each other, we used data only from male–female contexts to have certainty about the identity of the bird producing the warble (as the literature suggests warbles are produced by males) (Brockway, 1964). During the period of our study, we did not detect warble songs from females.

Analysis

We defined a warble as a vocal bout with a duration of at least 1 s, consisting of three or more warble elements (or notes, see below for more details on the note types). Notes separated by less than 1 s were considered to belong to a single warble bout, following definitions in the literature (Farabaugh et al., 1992; Tobin et al., 2019; Tu et al., 2011). When notes were very close together in time, we applied a separation of at least 10 ms and/or a dramatic change in note structure (see below for categories used to describe note structures) to identify the boundary between two notes. For each of the pair combinations described above, we analyzed 10 warble sequences (with a good signal:noise ratio and no background sounds), using the time stamps to ensure that these warbles occurred between 08:00 h and 18:00 h when the colonies were most active. One of the 18 pairs in the 2021 dataset emitted only a single warble in the 24 h duration, and was thus omitted from further analyses. Our final dataset consisted of a total of 340 warble sequences across the 2021 and 2022–23 recording experiments.

Next, we classified warble elements into different note types, broadly following the classification key employed by multiple published studies (Farabaugh et al., 1992; Tobin et al., 2019; Tu et al., 2011). Our dataset contained six such previously described categories, listed here according to the abbreviations used in Fig. 1B: ‘a’, alarm call-like elements; ‘b’, contact call-like elements (frequency modulated sounds); ‘c’, long harmonic calls (harmonic sounds with duration greater than 100 ms); ‘d’, short harmonic calls (harmonic sounds with duration less than 100 ms); ‘e’, ‘noisy’ calls that are broadband and non-harmonic; and ‘f’, clicks (extremely short broadband calls). In addition, our data contained an additional note type that we named ‘m’. These notes were generally low in amplitude, and did not fit into any of the above categories. We defined these as ‘soft calls’ (Fig. 1B; Fig. S1), on account of the low amplitude. Fig. S1 contains additional examples of each note, showing the range of note durations observed across each note type. Because multiple authors annotated note types, we performed cross-verification (where each author annotated the same warbles independently of the others) to quantify inter-observer agreement in note classification. Each person annotated five warbles from the first dataset (2021) and five from the second dataset (2022–23), and we then calculated the Levenshtein distances between each pair of warbles for each author (therefore, four values for each warble). Next, we normalized the Levenshtein distances to the length of each warble sequence to obtain the error as a percentage of sequence length. The average agreement in note type classification between observers was 80.16%, which is in line with values reported in the literature (Jones et al., 2001). We therefore deemed this an acceptable level of agreement to proceed with analysis of syntax.

Following the classification of notes, we quantified a number of metrics of warble structure and complexity, in order to compare them across contexts and colonies. First, we calculated the Shannon entropy and duration (in seconds) for each warble, using the pair combinations described above as our unit of measurement. The Shannon entropy (H) of a stochastic process with n states is defined as:
formula
(1)

Here, pi is defined as the probability of occurrence of the ith state and limpi→0pi·ln(pi)=0. The values of H can range from 0 to ln(n), with greater values for H indicating that the process is more unpredictable. In the case of budgerigar vocalizations, higher values of H indicate greater variation in the note composition of warble sequences, and thus greater complexity. We compared the Shannon entropy and duration values across contexts and across colonies (the latter for the 2021 data, and then again for the 2022–23 data across multiple pair combinations). Because there was only one male–male pair in colony 1, we only performed a statistical comparison of contexts for the data from colony 2. We recorded all possible pair combinations again in 2022–23 to compare the two colonies (and different contexts within colony 2) with each other before and after they were brought into contact. For these comparisons, we used linear mixed models where colony or context (depending on the comparison) was the fixed effect, and the identities of the two individuals were modelled as crossed random effects to ensure that individual identity did not pose a source of confounding variation. All models were run using the lme4 (Bates et al., 2015) package in R, and we tested statistical significance using the lmerTest (Kuznetsova et al., 2017) and afex (https://CRAN.R-project.org/package=afex) packages, which applied the Satterthwaite method. Lastly, we used the r.squaredGLMM function in the MuMIn package (https://CRAN.R-project.org/package=MuMIn) to estimate the marginal and conditional R2 values, as an estimate of effect size. The former estimates the proportion of variance explained by the fixed effect, and the latter the proportion of variation explained by the fixed and random effects combined.

Finally, we analyzed the syntactic structure of warbles, to compare how the composition of these songs varied across colonies before and after contact. Using the note classification detailed above, we first constructed note sequences for each warble. In addition to the above-described note types, and in order to incorporate the effects of silence into our analyses following a previous study (Bhat et al., 2022), we defined silent periods in the warble as instances where the inter-note interval exceeded 500 ms. This value of 500 ms was half of the duration between separate warble bouts (1 s, as defined above) and also represented a period greater than the duration of all observed note types, as described above. We labeled the silent periods in the warbles as ‘0’ in our note sequences.

To quantify whether certain notes tended to occur together in a warble sequence, we employed a co-occurrence metric that has been used for similar analyses in anuran vocal sequences (Bhat et al., 2022). We used this metric as it is free of assumptions about the underlying process generating vocal sequences in animal signals (for example, whether or not they follow Markov chain dynamics) (Kershenbaum et al., 2014). To elaborate, we defined dCij, as the probability that note type j occurs within a distance of d−1 notes of note type i. We computed dCij for warbles from each pair in our dataset. To compare these observed probabilities with those expected by chance, we next constructed artificial sequences, where warble elements were randomly distributed following a stationary distribution. Here, the probability of any note type occurring at any given position in the warble was given by the proportion of that note type in the warble sequences of that colony. We computed 50 such artificial sequences to have a robust estimate of dEij, where dEij is the probability that note type j occurs within a distance of d−1 notes of note type i due to random chance. We then computed dRij, defined as the ratio of dCij and dEij for warbles from individuals of each colony. dRij is, therefore, a measure of whether note type j occurs within a distance of d−1 notes from note i more or less often than expected by random chance. A value of dRij>1 suggests that note type j occurs within d−1 notes of note type i more often than expected by a random chance, and a value of dRij<1 suggests that note type j occurs within d−1 notes of note type i less often than expected by random chance. For the analysis described below, we used a d-value of 6, but repeated our analysis for d-values of 3 and 9 to ensure our choice of d-value did not influence our findings (Bhat et al., 2022).

For the analysis above, we computed the proportion of different note types in the warbles of individuals from the two colonies. Differences in the note proportions across colonies could occur for two reasons: (i) the note type occurred singly, but was emitted more or less often or (ii) a particular note type showed an increased or decreased tendency to repeat itself. These two possibilities result in very different outcomes in terms of the syntactic structure of sequences. To examine and distinguish between these possibilities, we computed the frequency of repeats (i.e. how often each note type occurred singly, as a repeat of two notes, and so on) for each note type in the warbles of each individual from both colonies. This enabled us to examine differences in the syntactic structure of warbles across colonies. To quantify statistically whether the two colonies exhibited syntactic differences greater than expected by chance (especially given the relatively low number of individuals in each colony), we performed a randomization test to examine colony-level differences. First, we randomly shuffled the colony labels of the warbles from the dataset obtained before contact and then calculated the frequency of different repeat lengths (ωd) of note type d in this randomized dataset (see Results). We then computed Cohen's d (a measure of effect size, or the distance between distributions) between the ωd of ‘colony 1’ warbles and ‘colony 2’ warbles for this randomized dataset. We repeated the above procedure 10,000 times to generate a random distribution of Cohen's d values for randomized colony 1 and colony 2, which gave us an expectation of repeat lengths under chance before contact between colonies. We performed the same randomization for sequences obtained after contact between the colonies. Finally, we computed Cohen's d between the observed ωd of colony 1 and colony 2 both before and after contact, and measured the Z-score of these observed values with respect to the two randomized distributions. This enabled us to test whether the observed differences in note repetition between the two colonies were greater than those expected by random chance, and what happened to these differences upon contact. A randomization analysis enabled us to examine whether differences in sequence structure (as measured by note repetitions) followed colony-specific patterns, independent of the individuals producing them and accounting for variation between individuals.

Warble duration and complexity do not differ across social contexts

On average, budgerigar warbles consisted of 83.66 notes (range: 8–954, total of 28,448 notes in our dataset across both 2021 and 2022–23 recording sessions, two males in colony 1 and three in colony 2). The results of linear mixed models are provided in Table S1. Warble duration and Shannon entropy did not exhibit a significant effect of context (male–male versus male–female either before or after contact; 0.001<marginal R2<0.03 for fixed effects across comparisons) (Fig. S2). As mentioned above, we were unable to perform a statistical comparison for colony 1 as there was only a single male–male pair in this colony. We next proceeded to examine whether warble structure (the information contained within a sequence) exhibited colony-specific patterns, as opposed to the context in which they were produced.

Warble complexity differs across colonies and converges upon contact

Before contact, differences in warble duration between the two colonies were not statistically significant (marginal R2=0.004, conditional R2=0.016). Warbles produced by individuals in colony 1, however, exhibited generally higher Shannon entropy values. This effect was statistically significant, but had a low effect size, with the fixed effects explaining 8% of variation (marginal R2=0.08, conditional R2=0.143, P<0.05) (Fig. 2A).

Fig. 2.

Box plots depicting the duration and Shannon entropy of warbles across the two colonies. Data are shown before (A) and after (B) contact. Shannon entropy H for the warbles of colony 1 differed significantly from that for the warbles of colony 2 before contact. The upper and lower edges of the box indicate the third and first quartile, respectively, the line inside each box indicates the median, and the whiskers represent 1.5 times the interquartile range; the diamonds above and below the boxes represent outliers.

Fig. 2.

Box plots depicting the duration and Shannon entropy of warbles across the two colonies. Data are shown before (A) and after (B) contact. Shannon entropy H for the warbles of colony 1 differed significantly from that for the warbles of colony 2 before contact. The upper and lower edges of the box indicate the third and first quartile, respectively, the line inside each box indicates the median, and the whiskers represent 1.5 times the interquartile range; the diamonds above and below the boxes represent outliers.

After the colonies were brought into contact and recorded again in 2022–23, however, the differences in Shannon entropy disappeared completely (duration: marginal R2<0.001, conditional R2=0.016; Shannon entropy: marginal R2<0.001, conditional R2=0.036; see Table S1 for more detailed results) (Fig. 2B). Because Shannon entropy is calculated using the proportions of different note types, our results suggest the possibility that the two colonies exhibited some minor differences in their use of different notes within the repertoire, and that these differences were lost upon contact between colonies. When we examined the relative proportions of different note types in the repertoire of each colony before contact, we observed differences between them (Fig. 3A), further supporting our assertion. Individuals in colony 2 produced a higher proportion of d notes in particular (see Fig. 1 for key). However, after contact, individuals from the two colonies produced similar proportions of d notes, as well as all other note types (Fig. 3B).

Fig. 3.

Pie chart depicting the proportions of different note types in the warbles of the two colonies. Data are shown before (A) and after (B) contact, as a proportion of the total notes produced. Note type d occurred in greater proportions in colony 2 before contact, but this difference was lost after the colonies were brought into contact. 0 here indicates periods of silence (see Materials and Methods for details).

Fig. 3.

Pie chart depicting the proportions of different note types in the warbles of the two colonies. Data are shown before (A) and after (B) contact, as a proportion of the total notes produced. Note type d occurred in greater proportions in colony 2 before contact, but this difference was lost after the colonies were brought into contact. 0 here indicates periods of silence (see Materials and Methods for details).

Colony signatures and syntactic convergence in budgerigar warbles

To examine whether differences in warble complexity between colonies, and the subsequent changes after contact, resulted from convergence of syntactic structure, we constructed note sequences for each warble. We then computed a co-occurrence matrix for warbles of each colony before and after contact. The ratio of observed to expected co-occurrence values (dRij) was >1 along the diagonal for both colonies (Fig. 4A). This suggests that all the note types had a propensity to repeat themselves within a warble sequence. Additionally, we found that the off-diagonal (between notes) dRij was <1. However, the tendency of d notes to repeat was higher in colony 2 than in colony 1. This, coupled with their higher proportion of occurrences in the colony 2 repertoire, suggested a higher-order signature of colony identity based on a tendency to repeat certain note types more often. The increased tendency for d notes to co-occur was lost after contact between colonies (Fig. 4B). This pattern was independent of the distance (d-value, see Materials and Methods) used in this analysis (Fig. S3).

Fig. 4.

Ratio of observed to expected co-occurrence (d=6) of different note types in the two colonies. Data are shown before (A) and after (B) contact. Redder or warmer colors in the ijth position in the matrix indicate that note type j occurred with note type i more often than expected by chance (see Materials and Methods for details). Note the higher propensity of note type d to co-occur with itself in colony 2 (A), indicating an increased probability of repetition, which was lost after the colonies were brought into contact (B).

Fig. 4.

Ratio of observed to expected co-occurrence (d=6) of different note types in the two colonies. Data are shown before (A) and after (B) contact. Redder or warmer colors in the ijth position in the matrix indicate that note type j occurred with note type i more often than expected by chance (see Materials and Methods for details). Note the higher propensity of note type d to co-occur with itself in colony 2 (A), indicating an increased probability of repetition, which was lost after the colonies were brought into contact (B).

We examined this pattern further by quantifying how often different note types repeated themselves within warble songs across the two colonies. In the warbles of colony 2, d notes exhibited an increased tendency to occur as repeats of 3 or more notes, and a decreased tendency to occur alone compared with colony 1 (Fig. 5A). Differences in d note repetitions between colony 1 and colony 2 were greater than expected by chance in a randomization test (d=0.63, Z=11.87, P<0.001; see Materials and Methods), consistent with non-random, colony-specific differences in syntax. After contact, however, individuals from the two colonies demonstrated broadly similar propensities to repeat all note types (Fig. 5B), and the differences between the two colonies were no longer significant in a randomization test, lying within the range expected by chance (d=0.1, Z=0.85, P=0.39). This pattern held true when considering each individual male in each colony, using data from male–female contexts; individual birds showed a pattern of note repetition specific to their colony of origin, which was lost after the colonies were brought into contact (Fig. S4). On the whole, the d note repetition patterns between the two colonies became more similar after contact (Fig. S4). These results were consistent whether grouping sequences according to colony or according to individual, suggesting that, at least by the metric of d note repetitions, there was relatively little individual variation within a colony. Thus, our data were broadly consistent with repetitions of different note types (d notes in this particular case) acting as signatures of colony identity embedded in a complex and lengthy warble. After colonies came into contact, syntactic convergence caused these colony signatures to disappear, such that individuals from both erstwhile colonies produced warbles of similar structure, duration and complexity.

Fig. 5.

Proportion of occurrences of note repetitions of different lengths, for each note type for each colony. Data are shown before (A) and after (B) contact. Insets depict repeat lengths of 2–7 (e.g. a certain note type occurring between two and seven times consecutively), enlarged for visual clarity. Males in colony 2 were more likely to repeat d notes before contact, but this higher-order dialectic variation disappeared after the colonies were brought into contact.

Fig. 5.

Proportion of occurrences of note repetitions of different lengths, for each note type for each colony. Data are shown before (A) and after (B) contact. Insets depict repeat lengths of 2–7 (e.g. a certain note type occurring between two and seven times consecutively), enlarged for visual clarity. Males in colony 2 were more likely to repeat d notes before contact, but this higher-order dialectic variation disappeared after the colonies were brought into contact.

Our study uncovers evidence that the warble song of budgerigars exhibits colony-specific syntactic structures (Figs 2 and 3). Secondly, we found that these differences in complexity resulted from an increased tendency of one colony to issue repetitive motifs of certain note types (Figs 4 and 5). This, in turn, led to differing proportions of notes in vocal sequences across colonies (Fig. 3), resulting in differences in complexity as measured by the Shannon entropy (Fig. 2). We thus hypothesize that the long, complex sequences of budgerigar warble songs putatively contain embedded higher-order signatures of dialectic variation. When colonies come into contact, these differences are lost rapidly. We found that after contact, the syntactic structure of warble sequences converged, and individuals from both erstwhile colonies sang warbles with similar syntactic structures. Although our colonies were relatively small (2 and 3 male birds, respectively) and we present data from a single colony unification event, the results were highly consistent across individual males within a colony (Fig. S4), with each bird following a colony-specific repetition pattern. A randomization analysis indicated that vocal sequences from the two colonies were more divergent than expected by chance before contact, but were statistically indistinguishable after contact. Although the d note repetition profiles in general demonstrated that birds from colony 2 tended to resemble those from colony 1 more after contact, we cannot say for certain whether the convergence occurred because birds from colony 2 learnt the song types produced by colony 1. This is because the song sequences are highly temporally complex, and the one feature that we identify here may potentially not capture all variation in song. However, it is possible that the birds from colony 1, which were acquired earlier, were socially dominant over the newer additions, and this represents an interesting direction for future study. Further, as our data only capture discrete time periods before and after contact, the dynamics of change in song are an interesting potential subject of study, especially as our observations suggest that change occurred in just a few months.

Syntax in animal sensory signals is both diverse and complex, with considerable variety in its form and function (Arnold and Zuberbühler, 2006; Backhouse et al., 2022; Bhat et al., 2022; Briefer et al., 2013; Engesser et al., 2016, 2019; Gentner et al., 2006; Isaac and Marler, 1963; Kershenbaum and Garland, 2015; Kershenbaum et al., 2012, 2016; Leroux et al., 2021; Ligon et al., 2018; Scholes, 2008). The structure of vocal sequences, particularly in vertebrates is known to convey information about context and behavioral state (Bhat et al., 2022; Ciaburri and Williams, 2019; Engesser et al., 2016; Leroux et al., 2021; Suzuki et al., 2018). For example, primates and pied babblers use different combinations of calls to convey different meanings (such as different behavioral states) (Arnold and Zuberbühler, 2006; Clay and Zuberbühler, 2011; Engesser and Townsend, 2019; Engesser et al., 2016, 2019) and chestnut-crowned babblers and hyraxes combine individually meaningless units into meaningful sequences that vary according to context (Engesser et al., 2019; Kershenbaum et al., 2012). In contrast, we found that the complexity of budgerigar vocal sequences did not vary according to social context (Fig. S2); that is, males produced similar warbles in both male–male and male–female contexts. Broadly, our results are concordant with another study suggesting that budgerigars sing similar warble songs to both male and female conspecifics (Tobin et al., 2019). Taken together, the available literature therefore suggests that warble song in budgerigars may serve a broader social function, as is observed in other social vertebrates (Candiotti et al., 2012; Deecke et al., 2010; Elowson and Snowdon, 1994; Mitani and Gros-Luis, 1998; Rendell and Whitehead, 2003).

The presence of unique, group-specific recognition signals may enable individuals to distinguish group members from non-group members (the password hypothesis) (Dahlin et al., 2014; Mammen and Nowicki, 1981; Tyack, 2008). Additionally, this similarity in call structure may also help mediate aggressive and affiliative interactions among group members (social association) (Dahlin et al., 2014; Mammen and Nowicki, 1981; Tyack, 2008). Most studies of parrots so far have suggested that these recognition signals are encoded in the acoustic structure of individual calls (Dahlin and Wright, 2012; Dahlin et al., 2014; Farabaugh et al., 1992; Hile and Striedter, 2000; Hile et al., 2000, 2005; Wanker and Fischer, 2001; Wanker et al., 2005). The structure of a flight call, for example, varies according to individual and colony (Dahlin et al., 2014; Farabaugh et al., 1992; Mammen and Nowicki, 1981). This acoustic structure (i.e. the time–frequency properties of an individual note) changes when individuals move between colonies or during pair bonding and maintenance (Dahlin et al., 2014; Farabaugh et al., 1992; Hile and Striedter, 2000; Hile et al., 2000, 2005). A similar pattern (social vocal accommodation) has been described for the acoustic structure of social calls in primates (Elowson and Snowdon, 1994; Zürcher et al., 2021). However, individual notes represent only a single dimension of the complexity of parrot acoustic communication. Parrots, being open-ended learners, frequently string together individual notes into complex vocal sequences that are also modified throughout their lifetime, often including mimicry of other sounds such as human speech (Balsby et al., 2017; Farabaugh et al., 1992; Pepperberg, 2009). Studies of intraspecific communication in parrots to date have focused almost exclusively on the acoustic properties of individual notes (Berg et al., 2012; Dahlin and Wright, 2009; Dahlin et al., 2014; Wright and Dahlin, 2017); very few studies have examined sequence structure or syntax in parrot vocalizations (Dahlin and Wright, 2009, 2012). We found that budgerigars from different social groups exhibit differences in the organization of notes in a warble sequence, and we suggest that the species broadly exhibits higher-order dialectic variation. Our data from colony unification presents preliminary evidence that higher-order variation is modified when groups come into contact, resulting in syntactic convergence (Figs 4 and 5). We therefore hypothesize that budgerigars possess complex, multiscale encoding of group identity in their acoustic signals, and can modify this by social learning throughout their lifetime as flocks associate and dissociate in the wild, similar to passerine birds such as starlings (Gentner et al., 2006). In nature, parrot social groups associate and dissociate into larger or smaller flocks in different seasons, and vocal convergence may help mediate social interactions as discussed above. Because our studies were undertaken in captivity, and represent a single colony unification event, we also highlight the need for further studies, preferably on wild parrots where individuals can be identified, to further examine the role that higher-order variation plays in social interactions.

Our quantitative examination of signal structure and note co-occurrence patterns presents an example of the strength of computational approaches in studying animal communication (Bhat et al., 2022; Isaac and Marler, 1963; Kershenbaum and Garland, 2015). Evolutionary history and social structure of vocalizing vertebrate taxa influence both the acoustic properties of individual notes and their arrangement into syntactic structures (Arato and Fitch, 2021; Wanker et al., 2005; Wright and Dahlin, 2017). Vocal learning also exerts a strong influence on this process, and can lead to rapid divergence or convergence in acoustic signals (Fishbein et al., 2020; Garland et al., 2011; Lachlan and Servedio, 2004; Noad et al., 2000; Wilkins et al., 2013; Yeh and Servedio, 2015). By presenting evidence of higher-order dialectic variation and syntactic convergence in budgerigars, our study suggests a need to examine dialectic variation and communication of identity at multiple scales of signal organization. Future studies may take advantage of unsupervised clustering and deep-learning approaches (Keen et al., 2021), which, as training datasets accumulate, will facilitate rapid, continuous assessment of long-term changes in song structure. We also underscore the possibilities presented by psittaciform birds as model systems to examine the influence of complex social systems and social vocal learning on the structure and temporal organization of vocal sequences.

We thank Ananda Shikhara Bhat for useful discussions and suggestions on analysis, Vaibhav Chhaya, Kunapareddy Kezia and Amey Danole for assistance during experiments, Prakash Raut and Suneel for bird care and Raghav Rajan and his lab for the use of their acoustic enclosures and for assistance and valuable discussions during the study. Finally, we thank Prof. Almut Kelber and two anonymous reviewers for their constructive feedback.

Author contributions

Conceptualization: A.J.M., A.K.; Methodology: A.J.M., N.W., P.B., V.A.; Software: A.J.M.; Validation: A.J.M., N.W., P.B.; Formal analysis: A.J.M., N.W., P.B.; Investigation: A.J.M., N.W., P.B.; Writing - original draft: A.J.M., A.K.; Writing - review & editing: A.J.M., A.K.; Supervision: A.K.; Project administration: A.K.; Funding acquisition: A.K.

Funding

A.K. is funded by an INSPIRE Faculty Award from the Department of Science and Technology, Government of India, an Early Career Research Grant (ECR/2017/001527) and a Core Research Grant (CRG/2022/000187) from the Science and Engineering Research Board (SERB), Government of India, as well as an institutional initiation grant from Indian Institute of Science Education and Research Bhopal. A.J.M. and N.W. are recipients of the KVPY Fellowship, and P.B. of the INSPIRE Fellowship from the Government of India.

Data availability

Base codes are available from GitHub: https://github.com/ab-madabhushi.

Allen
,
J. A.
,
Garland
,
E. C.
,
Dunlop
,
R. A.
and
Noad
,
M. J.
(
2018
).
Cultural revolutions reduce complexity in the songs of humpback whales
.
Proc. R. Soc. B Biol. Sci.
285
,
2
-
7
.
Arato
,
J.
and
Fitch
,
W. T.
(
2021
).
Phylogenetic signal in the vocalizations of vocal learning and vocal non-learning birds
.
Philos. Trans. R. Soc. B Biol. Sci.
376
,
20200241
.
Arnold
,
K.
and
Zuberbühler
,
K.
(
2006
).
Language evolution: semantic combinations in primate calls
.
Nature
441
,
303
.
Backhouse
,
F.
,
Dalziell
,
A. H.
,
Magrath
,
R. D.
and
Welbergen
,
J. A.
(
2022
).
Higher-order sequences of vocal mimicry performed by male Albert's lyrebirds are socially transmitted and enhance acoustic contrast
.
Proc. R. Soc. B Biol. Sci.
289
,
20212498
.
Balsby
,
T. J. S.
,
Eldermire
,
E. R. B.
,
Schnell
,
J. K.
,
Poesel
,
A.
,
Walsh
,
R. E.
and
Bradbury
,
J. W.
(
2017
).
Function of vocalization length and warble repertoire size in orange-fronted conures
.
Anim. Behav.
134
,
301
-
310
.
Bates
,
D.
,
Mächler
,
M.
,
Bolker
,
B.
and
Walker
,
S.
(
2015
).
Fitting linear mixed-effects models using lme4
.
J. Stat. Softw.
67
,
1
-
48
.
Berg
,
K. S.
,
Delgado
,
S.
,
Cortopassi
,
K. A.
,
Beissinger
,
S. R.
and
Bradbury
,
J. W.
(
2012
).
Vertical transmission of learned signatures in a wild parrot
.
Proc. R. Soc. B Biol. Sci.
279
,
585
-
591
.
Bhat
,
A. S.
,
Sane
,
V. A.
,
Seshadri
,
K. S.
and
Krishnan
,
A.
(
2022
).
Behavioural context shapes vocal sequences in two anuran species with different repertoire sizes
.
Anim. Behav.
184
,
111
-
129
.
Bohn
,
K. M.
,
Schmidt-French
,
B.
,
Schwartz
,
C.
,
Smotherman
,
M.
and
Pollak
,
G. D.
(
2009
).
Versatility and stereotypy of free-tailed bat songs
.
PLoS ONE
4
,
e6746
.
Bradbury
,
J. W.
and
Vehrencamp
,
S. L.
(
2011
).
Principles of Animal Communication
, 2nd edn.
Sinauer Associates
.
Briefer
,
E. F.
,
Rybak
,
F.
and
Aubin
,
T.
(
2013
).
Does true syntax or simple auditory object support the role of skylark song dialect?
Anim. Behav.
86
,
1131
-
1137
.
Brockway
,
B. F.
(
1964
).
Ethological studies of the budgerigar: reproductive behavior
.
Behaviour
23
,
294
-
323
.
Candiotti
,
A.
,
Zuberbühler
,
K.
and
Lemasson
,
A.
(
2012
).
Convergence and divergence in Diana monkey vocalizations
.
Biol. Lett.
8
,
382
-
385
.
Ciaburri
,
I.
and
Williams
,
H.
(
2019
).
Context-dependent variation of house finch song syntax
.
Anim. Behav.
147
,
33
-
42
.
Clay
,
Z.
and
Zuberbühler
,
K.
(
2011
).
Bonobos extract meaning from call sequences
.
PLoS ONE
6
,
e18786
.
Dahlin
,
C. R.
and
Wright
,
T. F.
(
2009
).
Duets in yellow-naped amazons: Variation in syntax, note composition and phonology at different levels of social organization
.
Ethology
115
,
857
-
871
.
Dahlin
,
C. R.
and
Wright
,
T. F.
(
2012
).
Does syntax contribute to the function of duets in a parrot, Amazona auropalliata?
Anim. Cogn.
15
,
647
-
656
.
Dahlin
,
C. R.
,
Young
,
A. M.
,
Cordier
,
B.
,
Mundry
,
R.
and
Wright
,
T. F.
(
2014
).
A test of multiple hypotheses for the function of call sharing in female budgerigars, Melopsittacus undulatus
.
Behav. Ecol. Sociobiol.
68
,
145
-
161
.
Dalziell
,
A. H.
and
Welbergen
,
J. A.
(
2016
).
Elaborate mimetic vocal displays by female superb lyrebirds
.
Front. Ecol. Evol.
4
,
1
-
13
.
Deecke
,
V. B.
,
Barrett-Lennard
,
L. G.
,
Spong
,
P.
and
Ford
,
J. K. B.
(
2010
).
The structure of stereotyped calls reflects kinship and social affiliation in resident killer whales (Orcinus orca)
.
Naturwissenschaften
97
,
513
-
518
.
Elowson
,
A. M.
and
Snowdon
,
C. T.
(
1994
).
Pygmy marmosets, Cebuella pygmaea, modify vocal structure in response to changed social environment
.
Anim. Behav.
47
,
1267
-
1277
.
Engesser
,
S.
and
Townsend
,
S. W.
(
2019
).
Combinatoriality in the vocal systems of nonhuman animals
.
Wiley Interdiscip. Rev. Cogn. Sci.
10
,
e1493
.
Engesser
,
S.
,
Ridley
,
A. R.
and
Townsend
,
S. W.
(
2016
).
Meaningful call combinations and compositional processing in the southern pied babbler
.
Proc. Natl. Acad. Sci. USA
113
,
5976
-
5981
.
Engesser
,
S.
,
Holub
,
J. L.
,
O'neill
,
L. G.
,
Russell
,
A. F.
and
Townsend
,
S. W.
(
2019
).
Chestnut-crowned babbler calls are composed of meaningless shared building blocks
.
Proc. Natl. Acad. Sci. USA
116
,
19579
-
19584
.
Farabaugh
,
S. M.
,
Brown
,
E. D.
and
Veltman
,
C. J.
(
1988
).
Song sharing in a group-living songbird, the Australian Magpie. Part II. Vocal sharing between territorial neighbors, within and between geographic regions, and between sexes
.
Behaviour
104
,
105
-
125
.
Farabaugh
,
S. M.
,
Brown
,
E. D.
and
Dooling
,
R. J.
(
1992
).
Analysis of warble song of the budgerigar Melopsittacus undulatus
.
Bioacoustics
4
,
111
-
130
.
Farabaugh
,
S. M.
,
Linzenbold
,
A.
and
Dooling
,
R. J.
(
1994
).
Vocal plasticity in budgerigars (Melopsittacus undulatus): evidence for social factors in the learning of contact calls
.
J. Comp. Psychol.
108
,
81
-
92
.
Fishbein
,
A. R.
,
Idsardi
,
W. J.
,
Ball
,
G. F.
and
Dooling
,
R. J.
(
2020
).
Sound sequences in birdsong: how much do birds really care?
Philos. Trans. R. Soc. B Biol. Sci.
375
,
20190044
.
Garland
,
E. C.
,
Goldizen
,
A. W.
,
Rekdahl
,
M. L.
,
Constantine
,
R.
,
Garrigue
,
C.
,
Hauser
,
N. D.
,
Poole
,
M. M.
,
Robbins
,
J.
and
Noad
,
M. J.
(
2011
).
Dynamic horizontal cultural transmission of humpback whale song at the ocean basin scale
.
Curr. Biol.
21
,
687
-
691
.
Gentner
,
T. Q.
,
Fenn
,
K. M.
,
Margoliash
,
D.
and
Nusbaum
,
H. C.
(
2006
).
Recursive syntactic pattern learning by songbirds
.
Nature
440
,
1204
-
1207
.
Hebets
,
E. A.
and
Papaj
,
D. R.
(
2005
).
Complex signal function: developing a framework of testable hypotheses
.
Behav. Ecol. Sociobiol.
57
,
197
-
214
.
Henry
,
L.
,
Barbu
,
S.
,
Lemasson
,
A.
and
Hausberger
,
M.
(
2015
).
Dialects in animals: evidence, development and potential functions
.
Anim. Behav. Cogn.
2
,
132
-
155
.
Hile
,
A. G.
and
Striedter
,
G. F.
(
2000
).
Call convergence within groups of female budgerigars (Melopsittacus undulatus)
.
Ethology
106
,
1105
-
1114
.
Hile
,
A. G.
,
Plummer
,
T. K.
and
Striedter
,
G. F.
(
2000
).
Male vocal imitation produces call convergence during pair bonding in budgerigars, Melopsittacus undulatus
.
Anim. Behav.
59
,
1209
-
1218
.
Hile
,
A. G.
,
Burley
,
N. T.
,
Coopersmith
,
C. B.
,
Foster
,
V. S.
and
Striedter
,
G. F.
(
2005
).
Effects of male vocal learning on female behavior in the budgerigar, Melopsittacus undulatus
.
Ethology
111
,
901
-
923
.
Hill
,
A. P.
,
Prince
,
P.
,
Snaddon
,
J. L.
,
Doncaster
,
C. P.
and
Rogers
,
A.
(
2019
).
AudioMoth: a low-cost acoustic device for monitoring biodiversity and the environment
.
HardwareX
6
,
e00073
.
Hultsch
,
H.
and
Todt
,
D.
(
1989
).
Memorization and reproduction of songs in nightingales (Luscinia megarhynchos): evidence for package formation
.
J. Comp. Physiol. A
165
,
197
-
203
.
Isaac
,
D.
and
Marler
,
P.
(
1963
).
Ordering of sequences of singing behaviour of mistle thrushes in relationship to timing
.
Anim. Behav.
11
,
179
-
188
.
Jones
,
A. E.
,
Ten Cate
,
C.
and
Bijleveld
,
C. C. J. H.
(
2001
).
The interobserver reliability of scoring sonagrams by eye: a study on methods, illustrated on zebra finch songs
.
Anim. Behav.
62
,
791
-
801
.
Keen
,
S. C.
,
Odom
,
K. J.
,
Webster
,
M. S.
,
Kohn
,
G. M.
,
Wright
,
T. F.
and
Araya-Salas
,
M.
(
2021
).
A machine learning approach for classifying and quantifying acoustic diversity
.
Methods Ecol. Evol.
12
,
1213
-
1225
.
Kershenbaum
,
A.
and
Garland
,
E. C.
(
2015
).
Quantifying similarity in animal vocal sequences: which metric performs best?
Methods Ecol. Evol.
6
,
1452
-
1461
.
Kershenbaum
,
A.
,
Ilany
,
A.
,
Blaustein
,
L.
and
Geffen
,
E.
(
2012
).
Syntactic structure and geographical dialects in the songs of male rock hyraxes
.
Proc. R. Soc. B Biol. Sci.
279
,
2974
-
2981
.
Kershenbaum
,
A.
,
Bowles
,
A. E.
,
Freeberg
,
T. M.
,
Jin
,
D. Z.
,
Lameira
,
A. R.
and
Bohn
,
K.
(
2014
).
Animal vocal sequences: Not the Markov chains we thought they were
.
Proc. R. Soc. B Biol. Sci.
281
.
Kershenbaum
,
A.
,
Blumstein
,
D. T.
,
Roch
,
M. A.
,
Akçay
,
Ç.
,
Backus
,
G.
,
Bee
,
M. A.
,
Bohn
,
K.
,
Cao
,
Y.
,
Carter
,
G.
,
Cäsar
,
C.
et al. 
(
2016
).
Acoustic sequences in non-human animals: a tutorial review and prospectus
.
Biol. Rev.
91
,
13
-
52
.
Kuznetsova
,
A.
,
Brockhoff
,
P. B.
and
Christensen
,
R. H.
(
2017
).
lmerTest package: tests in linear mixed effects models
.
J. Stat. Softw.
82
,
1
-
26
.
Lachlan
,
R. F.
and
Servedio
,
M. R.
(
2004
).
Song learning accelerates allopatric speciation
.
Evolution (N. Y.)
58
,
2049
-
2063
.
Leroux
,
M.
,
Bosshard
,
A. B.
,
Chandia
,
B.
,
Manser
,
A.
,
Zuberbühler
,
K.
and
Townsend
,
S. W.
(
2021
).
Chimpanzees combine pant hoots with food calls into larger structures
.
Anim. Behav.
179
,
41
-
50
.
Ligon
,
R. A.
,
Diaz
,
C. D.
,
Morano
,
J. L.
,
Troscianko
,
J.
,
Stevens
,
M.
,
Moskeland
,
A.
,
Laman
,
T. G.
and
Scholes
,
E.
(
2018
).
Evolution of correlated complexity in the radically different courtship signals of birds-of-paradise
.
PLoS Biol.
16
,
e2006962
.
Lipkind
,
D.
,
Zai
,
A. T.
,
Hanuschkin
,
A.
,
Marcus
,
G. F.
,
Tchernichovski
,
O.
and
Hahnloser
,
R. H. R.
(
2017
).
Songbirds work around computational complexity by learning song vocabulary independently of sequence
.
Nat. Commun.
8
,
1247
.
Mammen
,
D. L.
and
Nowicki
,
S.
(
1981
).
Individual differences and within-flock convergence in chickadee calls
.
Behav. Ecol. Sociobiol.
9
,
179
-
186
.
Marler
,
P.
and
Slabbekoorn
,
H.
(
2004
).
Nature's Music: the Science of Birdsong
.
London
,
UK
:
Elsevier Academic Press
.
Marler
,
P.
and
Tamura
,
M.
(
1962
).
Song “Dialects” in three populations of white-crowned sparrows
.
Condor
64
,
368
-
377
.
McComb
,
K.
,
Moss
,
C.
,
Sayialel
,
S.
and
Baker
,
L.
(
2000
).
Unusually extensive networks of vocal recognition in African elephants
.
Anim. Behav.
59
,
1103
-
1109
.
McCowan
,
B.
and
Hooper
,
S. L.
(
2002
).
Individual acoustic variation in Belding's ground squirrel alarm chirps in the High Sierra Nevada
.
J. Acoust. Soc. Am.
111
,
1157
-
1160
.
Mitani
,
J. C.
and
Gros-Luis
,
J.
(
1998
).
Chorusing and call convergence in chimpanzees: tests of three hypotheses
.
Behaviour
135
,
1041
-
1064
.
Mitoyen
,
C.
,
Quigley
,
C.
and
Fusani
,
L.
(
2019
).
Evolution and function of multimodal courtship displays
.
Ethology
125
,
503
-
515
.
Noad
,
M. J.
,
Cato
,
D. H.
,
Bryden
,
M. M.
,
Jenner
,
M.-N.
and
Jenner
,
K. C. S.
(
2000
).
Cultural revolution in whale songs
.
Nature
408
,
537
-
538
.
Pepperberg
,
I.
(
2009
).
Alex & me: How a Scientist and a Parrot Discovered a Hidden World of Animal Intelligence — and Formed a Deep Bond in the Process
.
New York
:
Harper Perennial
.
Peters
,
S. S.
,
Searcy
,
W. A.
and
Marler
,
P.
(
1980
).
Species song discrimination in choice experiments with territorial male swamp and song sparrows
.
Anim. Behav.
28
,
393
-
404
.
Rendell
,
L. E.
and
Whitehead
,
H.
(
2003
).
Vocal clans in sperm whales (Physeter macrocephalus)
.
Proc. R. Soc. B Biol. Sci.
270
,
225
-
231
.
Scholes
,
E.
III
. (
2008
).
Evolution of the courtship phenotype in the bird of paradise genus Parotia (Aves: Paradisaeidae): homology, phylogeny, and modularity
.
Biol. J. Linn. Soc.
94
,
491
-
504
.
Searcy
,
W. A.
,
Mcarthur
,
P. D.
,
Peters
,
S. S.
and
Marler
,
P.
(
1981
).
Response of male song and swamp sparrows to neighbour, stranger, and self songs
.
Behaviour
77
,
152
-
163
.
Suzuki
,
T. N.
,
Wheatcroft
,
D.
and
Griesser
,
M.
(
2018
).
Call combinations in birds and the evolution of compositional syntax
.
PLoS Biol.
16
,
e2006532
.
Tobin
,
C.
,
Medina-García
,
A.
,
Kohn
,
G. M.
and
Wright
,
T. F.
(
2019
).
Does audience affect the structure of warble song in budgerigars (Melopsittacus undulatus)?
Behav. Process.
163
,
81
-
90
.
Tu
,
H.-W.
,
Smith
,
E. W.
and
Dooling
,
R. J.
(
2011
).
Acoustic and perceptual categories of vocal elements in the warble song of budgerigars (Melopsittacus undulatus)
.
J. Comp. Psychol.
125
,
420
-
430
.
Tyack
,
P. L.
(
2008
).
Convergence of calls as animals form social bonds, active compensation for noisy communication channels, and the evolution of vocal learning in mammals
.
J. Comp. Psychol.
122
,
319
-
331
.
Wanker
,
R.
and
Fischer
,
J.
(
2001
).
Intra- and interindividual variation in the contact calls of spectacled parrotlets (Forpus conspicillatus)
.
Behaviour
138
,
709
-
726
.
Wanker
,
R.
,
Sugama
,
Y.
and
Prinage
,
S.
(
2005
).
Vocal labelling of family members in spectacled parrotlets, Forpus conspicillatus
.
Anim. Behav.
70
,
111
-
118
.
Wilkins
,
M. R.
,
Seddon
,
N.
and
Safran
,
R. J.
(
2013
).
Evolutionary divergence in acoustic signals: causes and consequences
.
Trends Ecol. Evol.
28
,
156
-
166
.
Wright
,
T. F.
and
Dahlin
,
C. R.
(
2017
).
Vocal dialects in parrots: Patterns and processes of cultural evolution
.
Emu
118
,
50
-
66
.
Yeh
,
D. J.
and
Servedio
,
M. R.
(
2015
).
Reproductive isolation with a learned trait in a structured population
.
Evolution (N. Y.)
69
,
1938
-
1947
.
Zdenek
,
C. N.
,
Heinsohn
,
R.
and
Langmore
,
N. E.
(
2015
).
Vocal complexity in the palm cockatoo (Probosciger aterrimus)
.
Bioacoustics
24
,
253
-
267
.
Zuberbühler
,
K.
(
2018
).
Combinatorial capacities in primates
.
Curr. Opin. Behav. Sci.
21
,
161
-
169
.
Zürcher
,
Y.
,
Willems
,
E. P.
and
Burkart
,
J. M.
(
2021
).
Trade-offs between vocal accommodation and individual recognisability in common marmoset vocalizations
.
Sci. Rep.
11
,
15683
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information