Access to this full-text is provided by PLOS.
Content available from PLOS One
This content is subject to copyright.
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 1 / 18
OPEN ACCESS
Citation: Pedruzzi L, Francesconi M, Galotti A,
Bogale BA, Palagi E, Lemasson A (2025) Wild
gelada monkeys detect emotional and prosocial
cues in vocal exchanges during aggression.
PLoS One 20(5): e0323295. https://doi.
org/10.1371/journal.pone.0323295
Editor: Teddy Lazebnik, Ariel University,
UNITED KINGDOM OF GREAT BRITAIN AND
NORTHERN IRELAND
Received: February 14, 2025
Accepted: April 4, 2025
Published: May 14, 2025
Copyright: © 2025 Pedruzzi et al
.
This is an
open access article distributed under the terms
of the Creative Commons Attribution License,
which permits unrestricted use, distribution,
and reproduction in any medium, provided the
original author and source are credited.
Data availability statement: All relevant
data are within the paper and its Supporting
Information files.
Funding: The research has been funded by
the Leakey Foundation (Science for recon-
ciliation: What an Ethiopian monkey tells
RESEARCH ARTICLE
Wild gelada monkeys detect emotional and
prosocial cues in vocal exchanges during
aggression
Luca Pedruzzi 1,2*, Martina Francesconi2, Alice Galotti2, Bezawork Afework Bogale3,
Elisabetta Palagi 2,4‡, Alban Lemasson1,5‡
1 EthoS (Ethologie Animale et Humaine) - U.M.R, Université de Rennes, Université de Normandie, CNRS,
Rennes, France, 2 Unit of Ethology, Department of Biology, University of Pisa, Pisa, Italy, 3 Department
of Zoological Sciences, College of Natural and Computational Sciences, Addis Ababa University, Addis
Ababa, Ethiopia, 4 Natural History Museum, University of Pisa, Calci, Pisa, Italy, 5 Institut Universitaire de
France, Paris Cedex, France,
‡ Shared senior authorship.
* luca.pedruzzi@phd.unipi.it, luca.pedruzzi@etudiant.univ-rennes.fr
Abstract
Recognizing vocal behaviours intended to benefit others is a crucial yet understudied
social skill. Primates with rich vocal repertoires and complex societies are excel-
lent models to track the evolution of such capacity. Here, we exposed wild gela-
das (Theropithecus gelada) to vocal exchanges between unfamiliar female victim
screams and male affiliative calls. The stimuli were arranged in sequences either
simulating vocal affiliation towards victims (scream-affiliative call) or violating such
order (affiliative call-scream), with varying emotional arousal conveyed by the affili-
ative call type. Measuring gazing activity towards the loudspeaker and the interrup-
tions of feeding, we show that monkeys were sensitive to the sequential order in
vocal exchanges as well as to the emotional arousal conveyed by affiliative calls. Our
field study suggests a prosocial use of vocalizations in wild monkeys and reveals that
foundational cognitive elements for processing vocal exchanges as meaningful third-
party interactions may have existed in our common ancestors with monkeys.
Introduction
Across human and non-human social animals, vocalizations play a crucial role in
facilitating appropriate responses to others’ needs [1–5]. In particular, they do so by
conveying a wide range of affective states through both within-call type variability,
such as changes in pitch, duration, or amplitude, and between-call type variability,
where different call categories signal distinct emotional or social contexts [1–5].
Vocal signals serve as vital cues for coordinating group activities, reinforcing social
bonds, and conveying intentions to help or cooperate, ultimately fostering group
cohesion [6]. Moreover, attending to a vocal exchange between two subjects can
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 2 / 18
inform about their relationship and the nature of the interaction (e.g., indirect social
information processing), especially in complex social systems where visual contact
can be constrained. By simply listening to call exchanges between two individuals,
several primates can indeed make inferences and obtain salient social information
about subject identity and directedness of the interaction [7], rank and kinship of the
interacting agents [8–10], and context of call production [11]. Playback experiments
show that some animal species demonstrate a certain degree of awareness of a
third-party conflict even when solely exposed to vocal cues [3,8,12,13] thus making
aggression and post-conflict behaviour excellent models to study vocal complexity
and prosocial tendencies [2,14].
In some cases, vocalizations can also actively elicit or function as prosocial
behaviours, as certain vocal signals may act as affiliative acts that promote social
bonding or solicit cooperative actions from others [14–20]. Prosociality is a funda-
mental building block of human and non-human tolerant societies [21] and com-
prises all those intentional behaviours expressed to benefit and help others [22]. The
immediate factors influencing prosocial acts are multifaceted, encompassing direct
requests from the beneficiary, potential benefits for the donor, and robust social
bonds between the interacting agents [23]. More in general, social (e.g., cooperative
alloparental breeding) and ecological (e.g., resource distribution) factors that charac-
terize different species play a key role in the evolution of prosocial and cooperative
behaviours [24]. While experimental studies suggest that some non-human animals
possess cognitive abilities to recognize prosocial from non-prosocial groupmates
based on past experience and social cues (primates, [25], corvids, [26]), scarce
data are present on animals’ ability to distinguish prosocial behaviours solely based
on vocal cues. For instance, dwarf mongooses (Helogale parvula) and marmosets
(Callithrix jacchus) have been shown to obtain information regarding the coopera-
tiveness of a subject from its vocalisations [27,28]. However, the two cases might be
explained by the cooperative breeding system of the two species, where individuals
rely on cooperative acts from group members; this could enhance the sensitivity to
social cues and might have selected fine strategies to punish or reward groupmates
according to their cooperation propensity [27,28].
In many primates, victims of aggression often emit screams, stereotyped noisy
drawn-out calls signalling the subject distress which can sometimes alert or recruit
prosocial responses and facilitate reconciliation and consolation [12,15–17,20,29].
Vocalisations produced by bystanders of aggression can sometimes act as a form of
affiliative and/or prosocial act towards victims, such as calming or comforting vocal-
izations that may facilitate social bonding [14,19,30,31]. Positive vocalizations, such
as coos, grunts, or other affiliative calls [17,19], can indeed trigger neuroendocrine
responses that promote social cohesion and reduce stress, mimicking the calming
effects of physical touch (humans, [32], rats, [33]). These vocal behaviours reflect
the complexity of primate communication, where vocalizations play an active role in
regulating social dynamics and facilitating cooperation [6].
Here, we use a playback paradigm with simulated vocal exchanges during
aggression in a wild population of geladas (Theropithecus gelada) to investigate
about peace-making, grant n° S202310431)
and by the following zoos and foundations
(funders of BRIDGES project, UNIPI, AOO
"BIO" - 0005878/2022), in alphabetic order:
Dudley Zoo (UK), Fondazione ARCA (Italy),
Giardino Zoologico di Pistoia (Italy), NaturZoo
Rheine (Germany), Parc des Félins (France),
Parco Natura Viva (Italy), Parco Zoo Falconara
(Italy), Rotterdam Zoo (The Netherlands),
Saint-Félicien (Canada), Wildlife Conservation
Benefit (Italy), Wilhelma Zoo (Germany), Zoo
de Cerza (France), Zoo de La Boissière du doré
(France). Rennes Métropole and the Région
Bretagne (France) covered the field expenses
for the first author while the University of
Pisa and University of Rennes funded student
fellowships. The funders had no role in study
design, data collection and analysis, decision to
publish, or preparation of the manuscript.
Competing interests: The authors declare no
competing interests.
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 3 / 18
their ability to extract social information from a third-party vocal interaction. Geladas, a monkey species endemic to
Ethiopia, live in complex multi-level societies (basic group units: One-Male Unit, OMU) with high levels of intra-unit
cohesion and inter-unit tolerance [34–36]. Conflict-resolution strategies adopted by the species involve both recon-
ciliation and spontaneous triadic affiliation towards victims [37,38]. Geladas show a richer vocal repertoire compared
to other phylogenetically close cercopithecines [29,39–42] and their socio-communicative complexity has often been
considered a precursor for that of humans [43,44]. For these reasons, among primates, geladas are optimal candidates
to study the recognition of compositional order and affective states in third-party vocal exchanges. Concerning affilia-
tive calls, not only do geladas produce grunts, low-intensity affiliative contact calls [45] commonly shared with Papio
baboons, but they have also evolved unique derived positive calls mostly produced by males such as moans and wob-
bles [29,41]. Moans are derived long drawn-out affiliative grunts (elongated version of a grunt) produced by both inha-
lation and exhalation [46]. Moans have been described as more salient and attractive than grunts for females [29,46],
possibly conveying higher emotional arousal [1,2]. The need for cross-sex bonding in their unique social system has
possibly driven gelada male vocal complexity [34], with evidence of control in the rhythm and melody during vocal
exchanges with females in emotionally aroused contexts [47]. Vocal sequences including moans are often produced
by leader males to their OMU’s females in contexts in which the need to maintain proximity is highest (e.g., moving,
presence of other group units) and often lead to grooming [34]. Data on the use of vocal signals during aggression are
relatively limited, and, more broadly, there is ongoing debate about whether vocalizations in non-human primates pri-
marily serve communicative functions or are largely involuntary expressions of emotions [48,49]. In geladas, although
intersexual communication has been studied in affiliative [34] and mating [40] contexts, research is lacking on male-
female vocal exchanges during negative contexts. Indeed, few studies only suggest that moans and grunts can be used
after aggression to affiliate with victims, either by the aggressors or by uninvolved third-party individuals [29,37,38],
highlighting a gap in the literature.
In the present study we use a field playback experiment with vocal exchanges between screams emitted by victims
of aggression and different affiliative vocal contacts produced by males (e.g., the aggressor or a third-party subject).
In particular, we employ a paradigm analogous to the violation of expectation, in which subjects are presented with
stimuli that defy their expectations, eliciting varying signs of surprise, vigilance, or attention, generally measured using
the gazing behaviour or behavioural interruptions as markers [50–52] (see section Video coding and statistical analy-
ses). These expectations are considered indicative of the mental principles the subjects hold about their surrounding
environment [53]. The method has been validated as an effective tool for uncovering the functional role and under-
lying rules governing primate vocal exchanges [50,51]. Specifically, we exposed the animals to acoustic sequences
either simulating vocal affiliation towards victims (when a victim’s scream was followed by an affiliative vocal contact,
scream-affiliative call) or violating such order (affiliative call-scream). Depending on the type of affiliative call used by
the male, we either have a low (grunts) or high (moans) emotional arousal conveyed by the call (Fig 1). We broadcast
calls from unfamiliar individuals to prevent any bias deriving from previous experience with socially bonded callers
[46]. This allows evaluating monkey ability to generally recognize the intrinsic nature of the message conveyed by the
signal. We posited that if geladas recognize the sequential order of a vocal exchange between the scream of a female
victim and the affiliative vocal contact of a male (Hypothesis A), we expect vocal exchanges violating a socially posi-
tive order (i.e., when a positive affective call is followed by a negative one) to attract more interest (in terms of gazing)
compared to vocal exchanges simulating vocal affiliation towards the victim (i.e., solved aggression) (Prediction A).
We expect monkeys to show some sort of sensitivity (shown by different levels of interest) also to the prosocial effort
expressed by the emotional arousal of the positive vocalisation emitted [2] (Hypothesis B). In this view, we predict
vocal exchanges containing male moans (i.e., moan-scream/scream-moan sequences), to elicit prolonged responses
than those containing male grunts (i.e., grunts-scream/scream-grunts sequences), affiliative calls expressing lower
arousal (Prediction B).
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 4 / 18
Materials and methods
Experimental model and subject details
The playback experiments were conducted between February and April 2024 in the unprotected area of Debre Libanos,
in the central highlands of Ethiopia (Oromia Region, Northwest Shewa zone) [54,55]. Specifically, 11 One-Male Units
(OMUs) were followed from 8:00 AM to 16:00 PM during the study period (for a total of 391 hours), as they daily visited
a large grazing area in Set Deber, one of the sites in Debre Libanos inhabited by gelada groups [55]. The study area (S1
Fig) allowed for optimal conditions for playback experiments (e.g., comparable microhabitat conditions across experi-
ments). We could test 10 fully-grown adult male subjects which were the alpha males of 10 of the 11 OMUs followed; the
subjects were individually recognized and familiar to human presence, thus allowing researchers to be in close proxim-
ity. The selection of adult male subjects was primarily driven by practical fieldwork considerations, including their easier
and unequivocal identification, the feasibility of consistent tracking, and their ability to reliably adhere to the experimental
paradigm and necessary procedural precautions (see Experimental Procedure). Moreover, adult males are the primary
producers of derived affiliative calls [34], playing a key role in regulating group dynamics. As a result, they are theoretically
the most attentive to variations in vocal sequence structure, making them particularly relevant for investigating sensitivity
to syntactic changes in our pilot playback experiment.
Fig 1. A study animal and examples of vocal exchanges used as stimuli. a) Picture of an adult gelada male in the study area (picture by EP); b)
graphical representations of the four conditions of simulated vocal exchanges used as stimuli during the experiments are represented, with four exam-
ples of spectrograms (obtained on Audacity© v. 3.3.2); c) Schematic representation of the different steps of the experimental procedure. See also S1 Fig.
https://doi.org/10.1371/journal.pone.0323295.g001
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 5 / 18
Ethical statement
Formal approval for the playback experiments was sought and obtained from the Bioethical Committee of the University of
Pisa (OPBA, n. 14/2023). Additionally, research procedures strictly adhered to the laws and approved guidelines set forth
by the Ethiopian Wildlife Conservation Authority (EWCA).
Stimuli preparation
To prepare the playback stimuli, LP and MF collected vocalisations produced by geladas living in the captive colony at
NaturZoo Rheine (Germany) in April-May 2023 [42]. Vocalisations were collected with a directional microphone (Senn-
heiser© MKE600) connected to a handy recorder (ZOOM H5©, sample rate: 44,100 Hz, resolution: 16-bit, wav format)
during spontaneous interactions (distance from the animals varied from 5 to 15 m). All subjects producing the stimuli (5
adult leader males, 40 females from the different OMUs; subjects individually recognized by LP and MF) were thus totally
unfamiliar to the study subjects. We selected three different call types among the gelada vocal repertoire: male grunts,
male exhale moans, and female screams [29]. These vocalisations can occur during aggression; from the observational
data we collected at NaturZoo Rheine, we recorded 103 agonistic interactions during which both moans/grunts and
scream were produced; in 58 cases the scream of a victim was followed by the affiliative call of a male, whereas in 45
events the affiliative call of the male preceded the scream of the female. During the recording phase, we also measured
the loudness (decibel, dB) of the three different vocalisations (mean ± SD = 59.1 ± 8.6 dB) with a professional sound meter
(SLM-25, Gain Express Holdings©). Only stimuli with a high sound-to-noise ratio were kept (e.g., no birds or overlap with
vocalisations from other geladas).
The recordings were then edited using Audacity© software (version 3.3.2) to create a pool of grunts (always pro-
duced in series, on average four in the present case), moans, and screams used to then build the final stimuli. All stimuli
contained one scream each but differed in the affiliative call (and relative order) in the sequence. To prevent any bias
in loudness variations, we normalized the amplitude of all calls and, before the beginning of the experiments and not
in the experimental area, we checked the speaker volume to reach about 60.0 Db [51,56] at the tested subject location
(15–20 metres, see Experimental procedure). To homogenize the amount of acoustic information conveyed by the two
positive call types, the duration of grunt series was adjusted to that of moans by repeating each grunt sequence twice
(totalizing 8 grunts vs 1 continuous moan). The pool of calls used to build stimuli thus comprised a total of 40 screams
(duration: mean ± SE = 5.214 ± 0.236 seconds), 20 grunt series (each composed of 8 exhale and inhaled grunts, duration:
mean ± SE = 5.579 ± 0.248 seconds), and 20 exhale moans (duration: mean ± SE = 5.372 ± 0.194 seconds). The five males
equally contributed to the calls recorded (grunt series and moans), preventing pseudo-replication. The calls were ran-
domly selected to build four different playback stimuli then randomly assigned to the 10 study subjects (for a total of 40
stimuli broadcast in 40 experimental sessions). The four stimuli to which each tested animal was exposed to contained
four affiliative call series produce by four captive males (identity randomly assigned) and four screams produced by four
captive females (identity randomly assigned) (Fig 1). A given vocalisation was thus never used twice during the entire
experimental period. The four stimuli were i) scream – grunt series (Sequential order; Type of affiliative call: Grunts), ii)
grunt series – scream (Sequential order; Type of affiliative call: Grunts) iii) scream – moan (Sequential order; Type of
affiliative call: Moan), iv) moan – scream (Sequential order; Type of affiliative call: Moan) (Fig 1). The silence duration from
the first to the second vocalisation comprised in each stimulus randomly ranged from 1.5–2.5 seconds, believed to be a
realistic latency time with which a male would react to a female scream. The mean duration of the total stimuli used was
10.2 seconds (SE: 0.6 s). For a given subject, stimuli duration did not differ either between the two conditions of Type of
affiliative call (grunts vs moan, Mann-Whitney-Wilcoxon Test, n1 = n2 = 20, W = 197, p = 0.95) or between the two conditions
of Sequential order (scream-affiliative call vs affiliative call-scream, Mann-Whitney-Wilcoxon Test, n1 = n2 = 20, W = 213,
p = 0.73). We checked for possible intra-call variability between experimental conditions. For a given subject, no difference
in the fundamental frequency (f0), generally indicating emotional arousal [1,4,57], was detected in the screams used in
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 6 / 18
the two conditions of Type of affiliative call (grunts vs moan, Mann-Whitney-Wilcoxon Test, n1 = n2 = 20, W = 208, p = 0.84)
and Sequential order (scream-affiliative call vs affiliative call-scream, Mann-Whitney-Wilcoxon Test, n1 = n2 = 20, W = 188.5,
p = 0.77). Similarly, the f0 of grunt series used in the two conditions of Sequential order did not differ (scream-affiliative call
vs affiliative call-scream, Mann-Whitney-Wilcoxon Test, n1 = n2 = 10, W = 70, p = 0.14), as well as that of moans in the two
conditions of Sequential order (scream-affiliative call vs affiliative call-scream, Mann-Whitney-Wilcoxon Test, n1 = n2 = 10,
W = 34, p = 0.25) [56].
Experimental procedure
Several precautions were taken to limit habituation and confounding factors [16,46,58]. The environmental conditions
and study subjects’ habituation to the researchers allowed the distance gelada-speaker to be kept as constant as possi-
ble during the different playback sessions (~15 metres). The speaker (MiPRO© MA-100 single channel Personal Wire-
less PA system) was positioned so that the sound would come from the direction where no other groups were present
not to simulate the vocal presence of an unfamiliar male in a known group. No aggression took place in the 30 minutes
before and no affiliative calls were audible by researchers in the 10 minutes before each playback. During experiments,
one observer (LP) played the stimulus (in.wav format) at distance via Bluetooth connected to the speaker hidden in
vegetation (S1b Fig). Another experimenter (MF, AG), generally visible to the tested gelada, video-recorded the animal
(SONY© handy-cam Full HD FDR-AX43A) in the 60 seconds before and after the sequence broadcast. After the exper-
imenter and the loudspeaker were set, we waited at least three minutes before starting the trial. This latter experiment-
er’s direction was of least 90° shifted compared to the direction of the acoustic stimulus. Each subject experienced the
four conditions in a random order (half sessions in the morning, 9–12am, half in the afternoon, 1–4pm). Subjects were
always tested during feeding, as the experimental area was a grazing area (S1 Fig) visited by the study groups mostly
for feeding, when not involved in social interactions, and with microhabitat visibility qualitatively similar across sessions.
We recorded whether other OMUs were in proximity (within ~ 50 metres) to the tested subject. The subject had to remain
visible and no event possibly affecting its vigilance (e.g., arrival of new group, aggression, vocalisations by other group
members) had to occur during the whole recording. No more than two playbacks were conducted per day and a condi-
tion was never played more than once per day. When a playback occurred, the identity of non-tested males who could
potentially hear the stimulus was coded so that we were sure that a minimum of 48 hours separated instances in which
a given gelada male could hear two stimuli. Moreover, to familiarize the animals with equipment and the presence of
researchers, situations comparable to “mock” experiments were conducted in which observers and equipment were
positioned as in actual experiments, but no stimulus was broadcast (the study groups were daily followed and recorded
during the whole experimental period).
Video coding and statistical analyses
Video recordings were analysed frame-by-frame (PotPlayer©, accuracy 0.02 sec), coding the following behaviours: i)
looking at the direction of the speaker, ii) self-directed behaviours (scratching, self-grooming, proxy for anxiety state
in primates [59]) (see S1 Table for definitions and operationalization). Every instance in which the tested subjects
turned their head towards the direction of the speaker was coded as looking at. Measures of the time subjects change
head orientation after playback are commonly used to evaluate their general interest in a given stimulus in non-human
primates [16,60]. Moreover, since geladas evolved a highly specialized graminivorous diet [61], their feeding implies
a seated position and the deployment of a series specialized hand movements [62] with their head facing the ground
[63], this allowed us to code for the iii) time in which the subject interrupted feeding (e.g., interrupting hand move-
ments and straightening up the head, only instances of feeding interruption for > 0.5 seconds were considered reliable
stops of the activity) as a possible proxy for vigilance state of the subject. Indeed, while vigilance seems relatively
cost free for upright feeders eating food that requires little manipulation [52], this is not the case for the gelada unique
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 7 / 18
feeding strategy [61–63]. To control for individual variations in vigilance or attentional state, behaviours were coded in
the 60 seconds before stimulus onset and in the 60 seconds after the whole stimulus broadcasting, thus after the ani-
mal could hear the entire sequence [46]. Behaviours were coded also before stimulus onset to measure the increase/
decrease of interest towards the speaker area after the stimulus compared to before [16,50]. No vocalisations by
the study subjects were recorded during the experimental sessions, neither before nor after stimulus presentation.
To evaluate the first interest towards the stimulus, we calculated the latency to return to the head position before the
stimulus (as in Pougnault et al. [50]), as subject started looking at the speaker before the end of the whole acous-
tic sequence. Then, to evaluate the prolonged response elicited by the stimulus, we measured the total time spent
gazing and calculated the mean duration per gaze, as well as measured the time spent in self-directed behaviours,
and in which they interrupted feeding (all measured in 60 seconds after – 60 seconds before stimulus broadcasting).
The coder (LP) was blind (muted videos labelled by MF) to the condition of the videos. Inter-observer reliability was
assessed with a second coder (AG) who coded 25% of videos blind to the conditions of the playback sessions and
who was in significant agreement with the first coder (for the four variables, Intraclass Correlation Coefficient (ICC) ≥
0.83, p < 0.001[64])
Model 1–5 – We ran five GLMMs with Duration of first interest to the loudspeaker (Model 1), Total gaze duration
towards the loudspeaker (Model 2), Mean duration per gaze to the loudspeaker (Model 3), Duration of stop in feeding
activity (Model 4), and Self-directed behaviours (Model 5) as response variables using a Gaussian distribution, log-
transforming the response variables for Model 1–3 after checking for model fit and diagnostics (DHARMa [65]). The
response variables are measured in tenths of seconds. Apart from the response variable, Model 1–5 were equally built:
the subject ID was included as random factor, whereas the fixed factors considered were the Sequential order (scream-
affiliative call/affiliative call-scream), the Type of affiliative call (grunts/moan) in the vocal exchange, the presence of Other
OMUs in proximity, and the Trial number.
All GLMMs allowed possible zero-inflation issues (zero values were present in all response variables) thanks to the
use of the glmmTMB package [66]; however, we did not model the zero-inflated part of the data including further model
components not to overcome optimal observations-predictors ratio in the models [67]. We checked for multicollinearity [67]
in the GLMMs using the ‘check_collinearity’ function (R package performance 0.4.4). ‘Low correlation’ was found for all the
fixed factors in the four models (VIF range: 1.06–1.65). We tested the models’ significance by comparing the full
with the control model (i.e., only including random and control factor(s)) through the Likelihood Ratio Test (LRT, Anova
with the ‘Chisq’ argument) and we then estimated the p-values of each predictor running LRTs between the full model
and the model not containing that predictor [68]. To check the models fit and possible overdispersion issues the pack-
age DHARMa 0.3.3.0 [65] was used. The GLMMs were not over-dispersed (dispersion range: 1.02–1.07, p-value range:
0.66–0.87), no outliers were detected (p-value range = 0.27–1), and normality of the residuals was confirmed via visual
inspection of Q-Q plots (Kolmogorov-Smirnov test, p-value range: 0.23–0.73). All the analyses were carried out using
Rstudio (http://www.r-project.org).
Results
A total of 40 experimental sessions were conducted on 10 adult gelada males, with each individual experiencing the four
different conditions, i) scream – grunt series, ii) grunt series – scream iii) scream – moan, iv) moan – scream; (Fig 1).
Model 1 – Duration of first interest towards the loudspeaker The full model was significantly different from the control one
(χ2
4 = 14.79, P = 0.005). The Sequential order and Type of affiliative call of the stimulus significantly affected the Duration
of first interest towards the loudspeaker, as the latency with which geladas turned their head back to the original position
after stimulus broadcasting was shorter when the stimulus had a sequential order simulating vocal affiliation to the victim
(Sequential order: | Coefficient|= 1.821, χ2 = 9.91, P = 0.001, Table 1, Fig 2a) and when the affiliative call in the stimulus had
higher emotional arousal (Type of affiliative call: | Coefficient|= 1.604, χ2 = 7.828, P = 0.005, Table 1).
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 8 / 18
Table 1. Duration of first interest to the loudspeaker.
Fixed Effects Coeff SE χ2df P
Model 1 - Duration of first interest to the loudspeaker
Intercept 2.540 0.948 – – –
Tested variables
Type of affiliative call (moans) 1.604 0.948 7.828 1 0.005
Sequential order (scream-affiliative call) -1.821 0.579 9.908 1 0.001
Trial number -0.304 0.254 1.434 1 0.231
Other OMUs in proximity (Yes) 0.671 0.660 1.034 1 0.309
Random factors: subject ID, Variance=1.357, SD=1.165.
Estimated parameters (Coeff), Standard Error (SE), and results of the Likelihood Ratio Tests (χ2) of Model 1. Significant P values are bold; df = degree(s)
of freedom; - = not applicable. Estimate ± SE refers to the difference of the response between the reported level of this categorical predictor and the refer-
ence category of the same predictor. nplaybacks
= 40 and nsubjects
= 10.
https://doi.org/10.1371/journal.pone.0323295.t001
Fig 2. The effect of stimulus sequential order on gelada gazing behaviour. Influence of the Sequential order of the simulated vocal exchange on a)
Duration of first interest towards the loudspeaker (measured in seconds) (Model 1: χ2 = 11.646, P < 0.001), b) Duration of total gazing towards the loud-
speaker (measured in seconds) (Model 2: χ2 = 17.36, P < 0.001), and c) Mean duration per gaze towards the loudspeaker (measured in seconds) (Model
3: χ2 = 14.718, P < 0.001). Subjects’ ID are represented with different colours and the Type of affiliative call of the stimulus is indicated by the point shape
(triangle = grunts; circle = moans). The boxes display the median value and first and third quartiles, whiskers are extended to the most extreme value
inside the 1.5-fold interquartile range.
https://doi.org/10.1371/journal.pone.0323295.g002
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 9 / 18
Model 2 – Duration of total gazing towards the loudspeaker (measured during 60 seconds after – during 60 seconds
before stimulus presentation, see Methods). The full model was significantly different from the null model (χ2
4 = 20.13,
P < 0.001). The Sequential order and the Type of affiliative call of the stimulus significantly affected the Duration of total
gazing towards the loudspeaker. Geladas looked for longer primarily when stimuli had the affiliative call-scream sequential
order (Sequential order: | Coefficient|= 2.695, χ2 = 18.10, P < 0.001, Table 2, Fig 2b) and secondarily when stimuli contained
a positive call expressing high emotional arousal (Type of affiliative call: | Coefficient|= 1.274, χ2 = 4.122, P = 0.042, Table 2,
Fig 3a).
Model 3 – Mean duration per gaze towards the loudspeaker (measured during 60 seconds after – during 60 seconds
before stimulus presentation, see Methods). The full model was significantly different from the null model (χ2
4 = 17.42,
P = 0.001). The Sequential order of the stimulus significantly affected the Mean duration per gaze towards the loud-
speaker, as geladas looked for longer towards the speaker when the stimulus violated the sequential order simulating
vocal affiliation to the victim (Sequential order: | Coefficient|= 1.778, χ2 = 19.569, P < 0.001, Table 3, Fig 2c). On the other
hand, the Type of affiliative call in the stimulus did not affect the Mean duration per gaze (Type of affiliative call: χ2 = 3.820,
P = 0.053, Table 3).
Model 4 – Duration of stop in feeding activity (measured during 60 seconds after – during 60 seconds before stimulus
presentation, see Methods). The full model significantly differed from the null one (χ2
4 = 13.62, P < 0.001). Geladas stopped
their feeding activity for longer when the affiliative call in the vocal exchange was a moan (Type of affiliative call: | Coef-
ficient|= 65.60, χ2 = 7.42, P = 0.006, Table 4, Fig 3b) and when the Sequential order of the stimulus did not simulate vocal
affiliation towards the victim (Sequential order: χ2 = 4.79, P = 0.029, Table 4).
Model 5 – Self-directed behaviours (60s after - 60s before stimulus) The full model did not significantly differ from the
control one (χ2
4 = 5.38, P = 0.25). None of the fixed factors included thus affected the variability of the time study subjects
spent in self-directed behaviours.
Discussion
To comprehend the evolutionary origins of the ability to extract salient social information from third-party vocal inter-
actions, it is essential to gather comparative data from species that exhibit social multilevel complexity and vocal rich-
ness. Geladas, with their intricate social structures and vocalization diversity, provide an ideal model for such studies.
Here, we exposed wild geladas to vocal interactions between unfamiliar female victim screams and male affiliative calls
with sequential order either simulating vocal affiliation towards the victim (scream-affiliative call) or violating such order
Table 2. Duration of total gazing towards the loudspeaker.
Fixed Effects Coeff SE χ2df P
Model 2 - Duration of total gazing towards the loudspeaker (60s after - 60s before stimulus)
Intercept 3.939 1.048 – – –
Tested variables
Type of affiliative call (moans) 1.274 0.627 4.122 1
1
0.042
Sequential order (scream-affiliative call) -2.695 0.633 18.098 1 <0.001
Trial number -0.529 0.279 3.604 1 0.058
Other OMUs in proximity (Yes) -0.141 0.720 0.038 1 0.844
Random factors: subject ID, Variance=1.774, SD=1.332.
Estimated parameters (Coeff), Standard Error (SE), and results of the Likelihood Ratio Tests (χ2) of Model 2. Significant P values are bold; df = degree(s)
of freedom; - = not applicable. Estimate ± SE refers to the difference of the response between the reported level of this categorical predictor and the refer-
ence category of the same predictor. nplaybacks
= 40 and nsubjects
= 10.
https://doi.org/10.1371/journal.pone.0323295.t002
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 10 / 18
(affiliative call-scream), using as affiliative call either grunt series (low emotional arousal) or moans (high emotional
arousal) (Fig 1). Geladas looked for longer towards the loudspeaker when the vocal exchange violated a possible conflict
resolution (Prediction A supported). Moreover, they also seemed sensitive to the affiliative call used towards victims as
study subjects interrupted feeding activity for longer and looked for longer towards the loudspeaker in response to vocal
exchanges containing affiliative moans compared to grunt series (Prediction B supported) (Fig 4).
A growing body of literature is uncovering how some primate species can compositionally combine calls to create new
meanings and vocal complexity [69–73]. Variability in communication can also be obtained by changing the order of calls
composing vocal sequences [69,74] as listeners can then attend to the referential changes induced by such permutations
[75]. The capacity of extracting meaning from the entire vocal sequences is crucial for survival in social groups [71]. This
Fig 3. The effect of the emotional arousal of the affiliative call in the sequence on gelada gazing behaviour and feeding interruption. Influence
of the Type of affiliative call of the stimulus (grunts vs moans) on a) Duration of total gazing towards the loudspeaker (seconds) (Model 2: χ2 = 3.994,
P = 0.046), b) Duration of stop in feeding activity (seconds) (Model 5: χ2 = 5.65, P = 0.017). Subjects’ ID are represented with different colours and the
Sequential order of the vocal exchange is indicated by the point shape (triangle = scream-affiliative call; circle = affiliative call-scream). The boxes display
the median value and first and third quartiles, whiskers are extended to the most extreme value inside the 1.5-fold interquartile range.
https://doi.org/10.1371/journal.pone.0323295.g003
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 11 / 18
has been so far studied in the vocal sequences produced by one subject, but such processing capacity might extend to
listeners attending to third-party vocal exchanges between two subjects. It is indeed known that primates respond differ-
ently to third-party vocal exchanges respecting vs not respecting different social rules [50,51,76]. Here, our results may
support this idea, suggesting that geladas possess sophisticated communicative abilities when processing third-party
vocal interactions.
Here, as the male vocally contacting the female could have been interpreted by the listeners as either the aggressor
or as a bystander, our results can be interpreted in two main ways. On the one hand, animals listening to the playbacks
might have interpreted them as reconciled vs not reconciled aggression and might have been sensitive to the resolution
of the conflict. On the other hand, they could have interpreted them as if a bystander was consoling a victim, demonstrat-
ing to be aware of a prosocial role of affiliative vocalizations. In any case, consistent with the species’ vocal and social
complexity, our findings suggest that geladas can extract meaningful social information (i.e., emotional arousal conveyed
to the victim of an aggression, presence of vocal affiliation and sequential outcome of a conflict) from acoustic cues and
might possibly distinguish between prosocial and non-prosocial vocal exchanges. Importantly, our observational data on
the natural occurrence of the vocal sequences used in our experiment (see Methods) indicate that both scream-affiliative
call and affiliative call-scream can occur during aggression in geladas; this suggests that the higher interest paid to stimuli
violating vocal affiliation towards a victim was not due to such order merely being unusual or absurd, but rather to the
sensitivity to the outcome of a conflict in a potentially nearby group unit.
Table 3. Mean duration per gaze towards the loudspeaker.
Fixed Effects Coeff SE χ2df P
Model 3 - Mean duration per gaze towards the loudspeaker (60s after - 60s before stimulus)
Intercept 2.158 0.646 – – –
Tested variables
Type of affiliative call (moans) 0.778 0.398 3.820 1 0.051
Sequential order (scream-affiliative call) -1.778 0.402 19.569 1 <0.001
Trial number -0.023 0.178 0.017 1 0.897
Other OMUs in proximity (Yes) 0.463 0.452 1.047 1 0.306
Estimated parameters (Coeff), Standard Error (SE), and results of the Likelihood Ratio Tests (χ2) of Model 3. Significant P values are bold; df = degree(s)
of freedom; - = not applicable. Estimate ± SE refers to the difference of the response between the reported level of this categorical predictor and the refer-
ence category of the same predictor. nplaybacks
= 40 and nsubjects
= 10.
https://doi.org/10.1371/journal.pone.0323295.t003
Table 4. Duration of stop in feeding activity.
Fixed Effects Coeff SE χ2df P
Model 5 - Duration of stop in feeding activity (60s after - 60s before stimulus)
Intercept 119.38 35.86 – – –
Tested variables
Type of affiliative call (moans) 65.60 24.08 7.421 1 0.006
Sequential order (scream-affiliative call) -53.13 24.28 4.789 1 0.029
Trial number -12.58 10.60 1.409 1 0.235
Other OMUs in proximity (Yes) -17.54 25.05 0.490 1 0.484
Random factors: subject ID, Variance=1.11*10-31, SD=3.25*10-16.
Estimated parameters (Coeff), Standard Error (SE), and results of the Likelihood Ratio Tests (χ2) of Model 5. Significant P values are bold; df = degree(s)
of freedom; - = not applicable. Estimate ± SE refers to the difference of the response between the reported level of this categorical predictor and the refer-
ence category of the same predictor. nplaybacks
= 40 and nsubjects = 10.
https://doi.org/10.1371/journal.pone.0323295.t004
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 12 / 18
Fig 4. Graphical abstract of the playback experiment. Geladas showed more interest towards stimuli violating a positive sequential order as well as
after stimuli with calls of high emotional arousal. Gelada monkeys seem to show cognitive ability to recognize vocal affiliation directed at victims.
https://doi.org/10.1371/journal.pone.0323295.g004
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 13 / 18
Our data can add important insights into the gelada complex social and vocal dynamics and, more in general, to social
eavesdropping of animals indirectly obtaining information about intra- and intergroup conflicts [13]. In species living in
complex social systems, such as geladas, acoustic cues can be especially valuable sources of social information. These
signals, produced during within- or between-group contests, provide bystanders with critical information about the out-
comes of others’ aggressive encounters [12]. This capacity goes beyond the previous experience of the subjects as, by
using unfamiliar stimuli, we show that geladas are able to generalize the actual social value associated with vocal interac-
tions. In particular, these primates might interpret call exchanges as meaningful third-party social interactions rather than
merely as a series of independent vocal events [28].
Importantly, even though future research is needed, this field experiment unveils the proximate causes leading to
prosocial behaviours after conflicts as well as that geladas might use vocal affiliation in reconciliation and consolation-like
behaviours towards victims [37,38]. Altogether, this suggests a greater complexity in the functional, intentional or expres-
sive, role of the species positive calls [29,34,45,46]. Vocalizations produced by receivers play indeed a crucial role in
inducing prosocial behaviours [15] and, when produced by donors of prosocial acts, can significantly influence the mam-
mal hormonal systems regulating social bonding, comparable to the effects of a physical contact [32,33]. Notably, here we
cannot draw conclusions on the processes underlying conflict-resolution strategies as well as on how the vocal exchanges
used are interpreted by listeners. Indeed, both human and non-human animal literature shows an ongoing debate on
whether prosocial behaviours such as affiliation provided to victims of aggression underlie sympathetic concern or more
self-protective and risk-mitigating reasons [77–79]. Moreover, although here we show that vocal exchanges in a certain
sequential order attract more interest by the tested animals, future research investigating the presence of social norms
[80] related to conflict-resolution strategies should investigate whether subjects are being rewarded/punished according to
their perceived propensity to cooperation, as it occurs in cooperative breeding species [28].
Geladas were also affected by the emotional arousal conveyed the male positive call in the vocal exchange. Indeed,
the study subjects seemed to remain more vigilant (i.e., stopped grazing behaviours, see Methods) and showed more
interest to the loudspeaker when the sequence broadcast contained moans, conveying higher emotional arousal com-
pared to grunts. Vocalisations are known to be optimal candidates for emotion transmission [81], as they show a wide
variability of between- and within-call-type features allowing to transmit a graded information about one’s affective state
[3–5]. The emotional nature (i.e., valence and arousal [82]) of different call variants can be adaptively (consciously or
not) discriminated by a recipient, who can differently react to the stimuli with variable “emotional content” [2]. Our experi-
ment indicates that male geladas also recognized the emotional arousal encoded by their different affiliative calls, in line
with evidence suggesting gelada females’ preference towards sequence containing a derived affiliative call [46]. Emotion
recognition has been considered either a prerequisite [2] or an indicator [83] for the occurrence of emotional contagion
based on the perception of vocalisations (for a review on issues in the study of empathy-related phenomena, see [77]).
Changes in the behavioural response of animals should indeed also be accompanied by changes in indicators of internal
affective states [84]. In our study, self-directed behaviours remained consistent across experimental conditions, effectively
ruling out the possibility that variations in anxiety levels [59,85] influenced by different stimuli could confound the subjects’
recognition abilities. Nevertheless, our result also suggests emotion and prosocial behaviour recognition do not systemati-
cally lead to emotional contagion [2,77,83], possibly implying that brain processes leading to mental representation can be
at play [75].
In this study, we demonstrate that wild monkeys exhibit different responses to vocal interactions depending on the
sequential order of calls, as well as their sensitivity to the positive call used in potential post-conflict interactions. It is
important to note that our results do not allow us to distinguish between cognitive and emotional processes that may
underlie their reactions. For example, we cannot conclusively determine if their responses are driven by higher-level
cognitive processes, such as reference-making, or by more basic emotional responses, such as emotional recognition.
Additionally, as is the case in most animal cognition experiments conducted in natural environments, our study is limited
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 14 / 18
by a small sample size. The wild setting in which this study took place provides a more ecologically valid environment but
also introduces variability due to individual differences, environmental conditions, and other factors that could influence
the results. Additionally, the stimuli used in this experiment were derived from a captive population, and we may expect
subtle inter-population differences in vocal production that could affect the generalizability of our findings.
In conclusion, this field experiment contributes to the growing body of evidence that animals use vocal cues from third-
party interactions as valuable sources of information. Additionally, it underscores that the selective pressures of complex
social environments may have favoured the development of cognitive abilities that represent evolutionary precursors to
certain aspects of human cognition. Our findings underscore the imperative ability of animals to quickly recognize the pos-
itive vs negative nature of communicative exchanges between conspecifics for their survival, aiming to better understand
how perception systems organize sensory information for rapid recognition. It opens new scenarios for future research
employing playback experiments in naturalistic conditions to investigate mental processing and social cognition in non-
human animals.
Supporting information
S1 Fig. The study site in Debre Libanos, Ethiopia. Pictures of a) the area where playback experiments were carried
out (Set Deber, Debre Libanos, Ethiopia; picture by LP) and b) the relative position of the study subject while feeding and
about to receive the stimulus and the experimenter (LP) hidden with the loudspeaker behind the vegetation and in a loca-
tion not visible to the tested animal (picture by EP).
(TIFF)
S1 Table. Ethogram of behaviours coded in the present study.
(DOCX)
S1 Dataset. Full raw dataset used for the statistical analyses of the study.
(XLSX)
Acknowledgments
We acknowledge the Ethiopian Wildlife Conservation Authority (EWCA) and the Oromia Forest and Wildlife Service for
granting us the permission to conduct our research, as well as Debre Libanos Woreda and Monastery for supporting our
stay in the community. Many thanks to Achim Johann (the coordinator of gelada European Endangered Species Pro-
gramme) for helping us in fundraising. Many thanks to Paolo Oliveri for the help in stimuli recording at NaturZoo Rheine
and to Sheleme Abiyou and Hailu Tilahun for their precious support in the field. We also thanks Birhanu and particularly
Kabebew Asefa Ylma for always taking care of us in the field. Finally, we wish to thank all the community and the kids
living in Set Deber for being respectful of our research and for spending great time with us: we will never forget your kind-
ness. Thanks to Fosca Mastrandrea for the drawing used in the Graphical abstract.
Author contributions
Conceptualization: Luca Pedruzzi, Elisabetta Palagi, Alban Lemasson.
Data curation: Luca Pedruzzi, Martina Francesconi, Alice Galotti, Elisabetta Palagi, Alban Lemasson.
Formal analysis: Luca Pedruzzi, Alban Lemasson.
Funding acquisition: Elisabetta Palagi.
Investigation: Luca Pedruzzi, Martina Francesconi, Alice Galotti, Elisabetta Palagi, Alban Lemasson.
Methodology: Luca Pedruzzi, Martina Francesconi, Alice Galotti, Bezawork Afework Bogale.
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 15 / 18
Project administration: Bezawork Afework Bogale, Elisabetta Palagi.
Supervision: Elisabetta Palagi, Alban Lemasson.
Visualization: Luca Pedruzzi.
Writing – original draft: Luca Pedruzzi.
Writing – review & editing: Luca Pedruzzi, Alice Galotti, Elisabetta Palagi, Alban Lemasson.
References
1. Briefer EF. Vocal expression of emotions in mammals: mechanisms of production and evidence. Journal of Zoology. 2012;288(1):1–20. https://doi.
org/10.1111/j.1469-7998.2012.00920.x
2. Briefer EF. Vocal contagion of emotions in non-human animals. Proc Biol Sci. 2018;285(1873):20172783. https://doi.org/10.1098/rspb.2017.2783
PMID: 29491174
3. Slocombe KE, Townsend SW, Zuberbühler K. Wild chimpanzees (Pan troglodytes schweinfurthii) distinguish between different scream types:
evidence from a playback study. Anim Cogn. 2009;12(3):441–9. https://doi.org/10.1007/s10071-008-0204-x PMID: 19112583
4. Lemasson A, Remeuf K, Rossard A, Zimmermann E. Cross-taxa similarities in affect-induced changes of vocal behavior and voice in arboreal mon-
keys. PLoS One. 2012;7(9):e45106. https://doi.org/10.1371/journal.pone.0045106 PMID: 22984618
5. Fischer J, Metz M, Cheney DL, Seyfarth RM. Baboon responses to graded bark variants. Animal Behaviour. 2001;61(5):925–31. https://doi.
org/10.1006/anbe.2000.1687
6. Briefer EF, Xie B, Engesser S, Sueur C, Freeberg TM, Brask JB. The power of sound: unravelling how acoustic communication shapes group
dynamics. Philos Trans R Soc Lond B Biol Sci. 2024;379(1905):20230182. https://doi.org/10.1098/rstb.2023.0182 PMID: 38768200
7. Engh AL, Hoffmeier RR, Cheney DL, Seyfarth RM. Who, me? Can baboons infer the target of vocalizations?. Animal Behaviour. 2006;71(2):381–7.
https://doi.org/10.1016/j.anbehav.2005.05.009
8. Bergman TJ, Beehner JC, Cheney DL, Seyfarth RM. Hierarchical classification by rank and kinship in baboons. Science. 2003;302(5648):1234–6.
https://doi.org/10.1126/science.1087513 PMID: 14615544
9. Borgeaud C, van de Waal E, Bshary R. Third-party ranks knowledge in wild vervet monkeys (Chlorocebus aethiops pygerythrus). PLoS One.
2013;8(3):e58562. https://doi.org/10.1371/journal.pone.0058562 PMID: 23520521
10. Cheney D, Seyfarth R. Recognition of other individuals’ social relationships by female baboons. Anim Behav. 1999;58(1):67–75. https://doi.
org/10.1006/anbe.1999.1131 PMID: 10413542
11. Arnold K, Zuberbühler K. Female Putty-Nosed Monkeys Use Experimentally Altered Contextual Information to Disambiguate the Cause of Male
Alarm Calls. PLoS One. 2013;8(6):2–8. https://doi.org/10.1371/journal.pone.0065660
12. Whitehouse J, Meunier H. An understanding of third-party friendships in a tolerant macaque. Sci Rep. 2020;10(1):9777. https://doi.org/10.1038/
s41598-020-66407-w PMID: 32555440
13. Morris-Drake A, Kern JM, Radford AN. Experimental evidence for delayed post-conflict management behaviour in wild dwarf mongooses. Elife.
2021;10:e69196. https://doi.org/10.7554/eLife.69196 PMID: 34725038
14. Wittig RM, Crockford C, Wikberg E, Seyfarth RM, Cheney DL. Kin-mediated reconciliation substitutes for direct reconciliation in female baboons.
Proc Biol Sci. 2007;274(1613):1109–15. https://doi.org/10.1098/rspb.2006.0203 PMID: 17301022
15. Heesen R, Austry DA, Upton Z, Clay Z. Flexible signalling strategies by victims mediate post-conflict interactions in bonobos. Philos Trans R Soc
Lond B Biol Sci. 2022;377(1860):20210310. https://doi.org/10.1098/rstb.2021.0310 PMID: 35934966
16. Lemasson A, Palombit RA, Jubin R. Friendships between males and lactating females in a free-ranging group of olive baboons (Papio hamadryas
anubis): evidence from playback experiments. Behav Ecol Sociobiol. 2007;62(6):1027–35. https://doi.org/10.1007/s00265-007-0530-z
17. Katsu N, Yamada K, Nakamichi M. Vocalizations during post-conflict affiliations from victims toward aggressors based on uncertainty in Japanese
macaques. PLoS One. 2017;12(5):e0178655. https://doi.org/10.1371/journal.pone.0178655 PMID: 28558070
18. Aere S, Pedruzzi L, Facondini G, Böye M, Palagi E, Lemasson A. The curious case of rhesus macaques: despotism does not prevent third-party
postconflict affiliation. Animal Behaviour. 2025;219:123022. https://doi.org/10.1016/j.anbehav.2024.10.036
19. Wittig RM, Crockford C, Seyfarth RM, Cheney DL. Vocal alliances in Chacma baboons (Papio hamadryas ursinus). Behav Ecol Sociobiol.
2007;61(6):899–909. https://doi.org/10.1007/s00265-006-0319-5
20. Fedurek P, Slocombe KE, Zuberbühler K. Chimpanzees communicate to two different audiences during aggressive interactions. Animal Behaviour.
2015;110:21–8. https://doi.org/10.1016/j.anbehav.2015.09.010
21. Kaplan HS, Hooper PL, Gurven M. The evolutionary and ecological roots of human social organization. Philos Trans R Soc Lond B Biol Sci.
2009;364(1533):3289–99. https://doi.org/10.1098/rstb.2009.0115 PMID: 19805435
22. Kopp KS, Kanngiesser P, Brügger RK, Daum MM, Gampe A, Köster M, et al. The proximate regulation of prosocial behaviour: towards a concep-
tual framework for comparative research. Anim Cogn. 2024;27(1):5. https://doi.org/10.1007/s10071-024-01846-w PMID: 38429436
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 16 / 18
23. Cronin KA. Prosocial behaviour in animals: the influence of social relationships, communication and rewards. Animal Behaviour. 2012;84(5):1085–
93. https://doi.org/10.1016/j.anbehav.2012.08.009
24. Burkart JM, Allon O, Amici F, Fichtel C, Finkenwirth C, Heschl A, et al. The evolutionary origin of human hyper-cooperation. Nat Commun.
2014;5:4747. https://doi.org/10.1038/ncomms5747 PMID: 25158760
25. Melis AP, Hare B, Tomasello M. Engineering cooperation in chimpanzees: tolerance constraints on cooperation. Animal Behaviour. 2006;72(2):275–
86. https://doi.org/10.1016/j.anbehav.2005.09.018
26. Seed AM, Clayton NS, Emery NJ. Cooperative problem solving in rooks (Corvus frugilegus). Proc Biol Sci. 2008;275(1641):1421–9. https://doi.
org/10.1098/rspb.2008.0111 PMID: 18364318
27. Kern JM, Radford AN. Experimental evidence for delayed contingent cooperation among wild dwarf mongooses. Proc Natl Acad Sci U S A.
2018;115(24):6255–60. https://doi.org/10.1073/pnas.1801000115 PMID: 29844179
28. Brügger RK, Willems EP, Burkart JM. Do marmosets understand others’ conversations? A thermography approach. Sci Adv. 2021;7(6):eabc8790.
https://doi.org/10.1126/sciadv.abc8790 PMID: 33536207
29. Gustison ML, le Roux A, Bergman TJ. Derived vocalizations of geladas (Theropithecus gelada) and the evolution of vocal complexity in primates.
Philos Trans R Soc Lond B Biol Sci. 2012;367(1597):1847–59. https://doi.org/10.1098/rstb.2011.0218 PMID: 22641823
30. Cheney DL, Seyfarth RM, Silk JB. The role of grunts in reconciling opponents and facilitating interactions among adult female baboons. Animal
Behaviour. 1995;50(1):249–57. https://doi.org/10.1006/anbe.1995.0237
31. Plotnik JM, de Waal FBM. Asian elephants (Elephas maximus) reassure others in distress. PeerJ. 2014;2:e278. https://doi.org/10.7717/peerj.278
PMID: 24688856
32. Seltzer LJ, Ziegler TE, Pollak SD. Social vocalizations can release oxytocin in humans. Proc Biol Sci. 2010;277(1694):2661–6. https://doi.
org/10.1098/rspb.2010.0567 PMID: 20462908
33. Seffer D, Schwarting RKW, Wöhr M. Pro-social ultrasonic communication in rats: insights from playback studies. J Neurosci Methods.
2014;234:73–81. https://doi.org/10.1016/j.jneumeth.2014.01.023 PMID: 24508146
34. Gustison ML, Tinsley Johnson E, Beehner JC, Bergman TJ. The social functions of complex vocal sequences in wild geladas. Behav Ecol Socio-
biol. 2019;73(1). https://doi.org/10.1007/s00265-018-2612-5
35. Pallante V, Ferrari PF, Gamba M, Palagi E. Embracing in a female-bonded monkey species (Theropithecus gelada). J Comp Psychol.
2019;133(4):442–51. https://doi.org/10.1037/com0000173 PMID: 30907610
36. Snyder-Mackler N, Beehner JC, Bergman TJ. Defining Higher Levels in the Multilevel Societies of Geladas (Theropithecus gelada). Int J Primatol.
2012;33(5):1054–68. https://doi.org/10.1007/s10764-012-9584-5
37. Palagi E, Leone A, Demuru E, Ferrari PF. High-Ranking Geladas Protect and Comfort Others After Conflicts. Sci Rep. 2018;8(1):15291. https://doi.
org/10.1038/s41598-018-33548-y PMID: 30327491
38. Leone A, Palagi E. Reconciling conflicts in a one-male society: the case of geladas (Theropithecus gelada). Primates. 2010;51(3):203–12. https://
doi.org/10.1007/s10329-010-0188-4 PMID: 20091206
39. Gustison ML, Bergman TJ. Divergent acoustic properties of gelada and baboon vocalizations and their implications for the evolution of human
speech. J Lang Evol. 2017;2(1):20–36. https://doi.org/10.1093/jole/lzx015 PMID: 31402984
40. Zanoli A, Gamba M, Lemasson A, Norscia I, Palagi E. Inter-sexual multimodal communication during mating in wild geladas: the leading role of
females. Curr Zool. 2021;68(5):570–80. https://doi.org/10.1093/cz/zoab093 PMID: 36324533
41. Aich H, Moos-Heilen R, Zimmermann E. Vocalizations of adult gelada baboons (Theropithecus gelada): acoustic structure and behavioural context.
Folia Primatol (Basel). 1990;55(3–4):109–32. https://doi.org/10.1159/000156508 PMID: 2262173
42. Pedruzzi L, Francesconi M, Palagi E, Lemasson A. The sound of yawns makes geladas yawn. Sci Rep. 2024;14(1):361. https://doi.org/10.1038/
s41598-023-49797-5 PMID: 38185686
43. Gustison ML, Semple S, Ferrer-I-Cancho R, Bergman TJ. Gelada vocal sequences follow Menzerath’s linguistic law. Proc Natl Acad Sci U S A.
2016;113(19):E2750-8. https://doi.org/10.1073/pnas.1522072113 PMID: 27091968
44. Bergman TJ. Speech-like vocalized lip-smacking in geladas. Curr Biol. 2013;23(7):R268-9. https://doi.org/10.1016/j.cub.2013.02.038 PMID:
23578870
45. Painter MC, Gustison ML, Snyder-Mackler N, Tinsley Johnson E, le Roux A, Bergman TJ. Acoustic variation and group level convergence of
gelada, Theropithecus gelada, contact calls. Animal Behaviour. 2024;207:235–46. https://doi.org/10.1016/j.anbehav.2023.10.002
46. Gustison ML, Bergman TJ. Vocal complexity influences female responses to gelada male calls. Sci Rep. 2016;6:19680. https://doi.org/10.1038/
srep19680 PMID: 26790770
47. Richman B. Rhythm and melody in gelada vocal exchanges. Primates. 1987;28(2):199–223. https://doi.org/10.1007/bf02382570
48. Fröhlich M, Sievers C, Townsend SW, Gruber T, van Schaik CP. Multimodal communication and language origins: integrating gestures and vocal-
izations. Biol Rev Camb Philos Soc. 2019;94(5):1809–29. https://doi.org/10.1111/brv.12535 PMID: 31250542
49. Gruber T, Grandjean D. A comparative neurological approach to emotional expressions in primate vocalizations. Neuroscience & Biobehavioral
Reviews. 2017; 73:182–90. https://doi.org/10.1016/j.neubiorev.2016.12.004
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 17 / 18
50. Pougnault L, Levréro F, Mulot B, Lemasson A. Breaking conversational rules matters to captive gorillas: A playback experiment. Sci Rep. 2020;
10(1):1–10. https://doi.org/10.1038/s41598-020-63923-7
51. Bouchet H, Koda H, Lemasson A. Age-dependent change in attention paid to vocal exchange rules in Japanese macaques. Animal Behaviour.
2017;129:81–92. https://doi.org/10.1016/j.anbehav.2017.05.012
52. Allan ATL, Hill RA. What have we been looking at? A call for consistency in studies of primate vigilance. Am J Phys Anthropol. 2018;165 Suppl
65:4–22. https://doi.org/10.1002/ajpa.23381 PMID: 29380883
53. Ginnobili S, Olmos AS. Empirical assumptions behind the violation of expectation experiments in human and non-human animals. Hist Philos Life
Sci. 2021;43(3):106. https://doi.org/10.1007/s40656-021-00459-7 PMID: 34462865
54. Abie K, Bekele A. Population Estimate, Group Size and Age Structure of the Gelada Baboon (Theropithecus Gelada) around Debre-Libanos,
Northwest Shewa Zone, Ethiopia. Global Journal of Science Frontier Research 2017; 17(1):27–33.
55. Galotti A, Francesconi M, Pedruzzi L, Abiyou Gamessa S, Lemasson A, Bogale BA, et al. Bridging Species Divides: Affiliative Interactions
between Theropithecus gelada and Papio anubis in an Unprotected Area in Ethiopia. Int J Primatol. 2024;46(1):4–8. https://doi.org/10.1007/
s10764-024-00450-6
56. Leroux M, Hetem RS, Hausberger M, Lemasson A. Cheetahs discriminate familiar and unfamiliar human voices. Sci Rep. 2018;8(1):15516. https://
doi.org/10.1038/s41598-018-33971-1 PMID: 30341369
57. Russell JA, Bachorowski J-A, Fernandez-Dols J-M. Facial and vocal expressions of emotion. Annu Rev Psychol. 2003;54:329–49. https://doi.
org/10.1146/annurev.psych.54.101601.145102 PMID: 12415074
58. Fischer J, Noser R, Hammerschmidt K. Bioacoustic field research: a primer to acoustic analyses and playback experiments with primates. Am J
Primatol. 2013;75(7):643–63. https://doi.org/10.1002/ajp.22153 PMID: 23592340
59. Maestripieri D, Schino G, Aureli F, Troisi A. A modest proposal: displacement activities as an indicator of emotions in primates. Animal Behaviour.
1992;44(5):967–79. https://doi.org/10.1016/s0003-3472(05)80592-5
60. Leroux M, Schel AM, Wilke C, Chandia B, Zuberbühler K, Slocombe KE, et al. Call combinations and compositional processing in wild chimpan-
zees. Nat Commun. 2023;14(1):2225. https://doi.org/10.1038/s41467-023-37816-y PMID: 37142584
61. Dunbar RIM, Bose U. Adaptation to grass-eating in gelada baboons. Primates. 1991;32(1):1–7. https://doi.org/10.1007/bf02381596
62. Truppa V, Gamba M, Togliatto R, Caselli M, Zanoli A, Palagi E, et al. Manual preference, performance, and dexterity for bimanual grass-feeding
behavior in wild geladas (Theropithecus gelada). Am J Primatol. 2024;86(5):e23602. https://doi.org/10.1002/ajp.23602 PMID: 38299312
63. Dunbar RIM. Feeding Ecology of Gelada Baboons: a Preliminary Report. In: Clutton-Brock TH, Editor. Primate Ecology: Studies of Feeding and
Ranging Behavior in Lemurs, Monkey and Apes. Academic Press; 1977; 251–73.
64. Koo TK, Li MY. A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research. J Chiropr Med. 2016;15(2):155–
63. https://doi.org/10.1016/j.jcm.2016.02.012 PMID: 27330520
65. HartigFDHARMa: residual diagnostics for hierarchical (multi-level/mixed) regression modelsR package version 0.4.62020https://CRAN.R-project.
org/package=DHARMa
66. Brooks ME, Kristensen K, van Benthem KJ, Magnusson A, Berg CW, Nielsen A, et al. glmmTMB balances speed and flexibility among packages
for zero-inflated generalized linear mixed modeling. R J. 2017;9(2):378–400.
67. Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens MHH, et al. Generalized linear mixed models: a practical guide for ecology
and evolution. Trends Ecol Evol. 2009;24(3):127–35. https://doi.org/10.1016/j.tree.2008.10.008 PMID: 19185386
68. Dobson AJ, Barnett AG. An Introduction to Generalized Linear Models. 4th ed. CRC Press; 2018. p. 392.
69. Berthet M, Mesbahi G, Pajot A, Cäsar C, Neumann C, Zuberbühler K. Titi monkeys combine alarm calls to create probabilistic meaning. Science
Advances. 2019;5(5):1–9. https://doi.org/10.1126/sciadv.aav3991
70. Clay Z, Zuberbühler K. Bonobos extract meaning from call sequences. PLoS One. 2011;6(4):e18786. https://doi.org/10.1371/journal.pone.0018786
PMID: 21556149
71. Ouattara K, Lemasson A, Zuberbühler K. Campbell’s monkeys use affixation to alter call meaning. PLoS One. 2009;4(11):e7808. https://doi.
org/10.1371/journal.pone.0007808 PMID: 19915663
72. Candiotti A, Zuberbühler K, Lemasson A. Context-related call combinations in female Diana monkeys. Anim Cogn. 2012;15(3):327–39. https://doi.
org/10.1007/s10071-011-0456-8 PMID: 21947942
73. Zuberbühler K, Lemasson A. Primate Communication: Meaning from Strings of Calls. In: Language and Recursion. New York, NY: Springer New
York; 2014; p. 115–25.
74. Girard-Buttoz C, Bortolato T, Laporte M, Grampp M, Zuberbühler K, Wittig RM, et al. Population-specific call order in chimpanzee greeting vocal
sequences. iScience. 2022;25(9):104851. https://doi.org/10.1016/j.isci.2022.104851 PMID: 36034222
75. Zuberbühler K, Bickel B. Transition to language: From agent perception to event representation. Wiley Interdiscip Rev Cogn Sci. 2022;13(6):e1594.
https://doi.org/10.1002/wcs.1594 PMID: 35639563
76. Lemasson A, Glas L, Barbu S, Lacroix A, Guilloux M, Remeuf K, et al Youngsters do not pay attention to conversational rules: Is this so for nonhu-
man primates? Scientific Reports 2011 1:12–5. https://doi.org/10.1038/srep00022
PLOS One | https://doi.org/10.1371/journal.pone.0323295 May 14, 2025 18 / 18
77. Adriaense JEC, Koski SE, Huber L, Lamm C. Challenges in the comparative study of empathy and related phenomena in animals. Neurosci Biobe-
hav Rev. 2020;112:62–82. https://doi.org/10.1016/j.neubiorev.2020.01.021 PMID: 32001272
78. Yamamoto S. Primate empathy: three factors and their combinations for empathy-related phenomena. Wiley Interdiscip Rev Cogn Sci.
2017;8(3):10.1002/wcs.1431. https://doi.org/10.1002/wcs.1431 PMID: 27977913
79. Koski SE, Sterck EHM. Triadic postconflict affiliation in captive chimpanzees: does consolation console?. Animal Behaviour. 2007;73(1):133–42.
https://doi.org/10.1016/j.anbehav.2006.04.009
80. Westra E, Fitzpatrick S, Brosnan SF, Gruber T, Hobaiter C, Hopper LM, et al. In search of animal normativity: a framework for studying social
norms in non-human animals. Biol Rev Camb Philos Soc. 2024;99(3):1058–74. https://doi.org/10.1111/brv.13056 PMID: 38268182
81. Owren MJ, Rendall D. An affect-conditioning model of nonhuman primate vocal signaling. In: Communication. Springer, Boston, MA; 1997. pp.
299–346.
82. Paul ES, Sher S, Tamietto M, Winkielman P, Mendl MT. Towards a comparative science of emotion: Affect and consciousness in humans and
animals. Neurosci Biobehav Rev. 2020;108:749–70. https://doi.org/10.1016/j.neubiorev.2019.11.014 PMID: 31778680
83. Nieuwburg EGI, Ploeger A, Kret ME. Emotion recognition in nonhuman primates: How experimental research can contribute to a better under-
standing of underlying mechanisms. Neurosci Biobehav Rev. 2021;123:24–47. https://doi.org/10.1016/j.neubiorev.2020.11.029 PMID: 33453306
84. Huber A, Barber ALA, Faragó T, Müller CA, Huber L. Investigating emotional contagion in dogs (Canis familiaris) to emotional sounds of humans
and conspecifics. Anim Cogn. 2017;20(4):703–15. https://doi.org/10.1007/s10071-017-1092-8 PMID: 28432495
85. Duboscq J, Romano V, Sueur C, MacIntosh AJJ. Scratch that itch: revisiting links between self-directed behaviour and parasitological, social and
environmental factors in a free-ranging primate. R Soc Open Sci. 2016;3(11):160571. https://doi.org/10.1098/rsos.160571 PMID: 28018646
