The “Ooooooohhhh!” a human being cries out when they stub their toe might sound a pretty similar to the “Ooooooohhhh!” they cry out at the end of their mating ritual, but they two calls are different. An important part of human-to-human communication is our ability to extract information from context-specific calls and integrate it with other information we already have to make sense of what we’re hearing. It’s how we know, if we’re standing in one room and the TV is on in another, the difference between the scream of a serial killer’s victim in a slasher movie and the scream of a hero going into battle in an action blockbuster. We might not know what kind of movie is on in there, but we can at least identify which end of a blade the screamer might be on.
Katie Slocombe, a lecturer at the University of York’s psychology department, has spent her career tracing the evolution of different aspects of human language. More often than not, she finds herself starting with pants, grunts, hoots and hollers of chimpanzees. Many people find this surprising, Slocombe has said, but they shouldn’t. Finding an evolutionary explanation for any part of human language is difficult. Unlike, say, wrist bones, spoken language hasn’t left any fossil remains behind for us to study. Genetic evidence from our hominid ancestors suggests that we evolved our capacity for complex spoken language in a very short window of time, so it’s likely that the cognitive abilities underlying language emerged farther back in the primate lineage. Hence it makes perfect sense to look to other living primates, apes and monkeys, for clues to language’s origins.
Chimpanzees, our closest living relatives, produce specific screams when locked in confrontation with each other. They vary their screams depending on their social role in a fight, with victims and aggressors producing acoustically distinct screams, and the victims additionally varying the structure of their screams based on amount of aggression they’re facing. Many things a listener would want to know about the conflict and the chimps involved is encoded in the sounds of the fight. Research has shown that chimps can discriminate the differences between these varying calls, and Slocombe wondered if they also share our ability to pull meaningful social information from them and make inferences about conflicts they can’t see.
To find their answer, Slocombe and her colleagues, Tanja Kaller, Josep Call and Klaus Zuberbühler, paid a visit to the chimpanzees living at the Wolfgang Köhler Primate Research Centre (WKPRC), Leipzig, Germany. They recorded screams from naturally occurring conflicts within the troop and monitored the responses of several bystander chimps to two types of scream sequences. One, the congruent sequence, consisted of calls that were in accordance with existing social dominance relations (that is, dominant animals were the aggressors and lower ranking ones were the victims), and the other, the incongruent sequence, consisted of calls that violated the hierarchy.
The researchers hypothesized that if chimps could discriminate and figure out the meaning of the different calls and the socialconstraints under which the two callers live, they would respond more to the latter sequences (in line with results from studies of other animals and human infants). If they couldn’t understand they context of the calls, their responses should be random, or in the other direction since, since the congruent sequences are more acoustically interesting.
The call of the wild
The incongruent call sequence consisted of a low-ranking chimp giving an aggressor scream, followed by a higher-ranking chimp giving a victim scream. This is an unusual event, because chimpanzees are rarely pushed around by lower-ranking group members. Congruent sequence would logically seem to consist of an inverted sequence, a high-ranking aggressor scream then low-ranking victim scream, but juxtaposing this with the incongruent sequence presented the problem of novelty. Because the incongruent sequence was so unusual, interpreting a strong response to it would be difficult. Did the bystander chimps respond because they could only understand the conflict by extracting social information from the calls, or simply because it was an unusual thing to hear?
To solve the problem, the researchers needed a workaround so that the sequences of aggressor and victim screams remained identical in the two conditions. For the congruent sequence, they reused the incongruent scream sequence, but added a third voice to the mix. The “pant-hoot” of a top-ranking male was slipped into the middle of the recording so that it overlapped with parts of the aggressor and the victim screams. Two of the screams were the same as the incongruent sequence, taking away from its novelty, and the third voice made the scenario socially plausible: it sounds like the high-ranking victim’s scream was elicited by the dominant male, rather than the low-ranking individual.
Smile, You’re on Candid Camera
Three males and seven females (10-31 years old) from the troop of 18 housed at the research center participated in the experiment. Each round involved 6-7 of the chimps as the one listening subject, two or three call providers or two “extras.” Assuming that the social chimps kept track of each other’s whereabouts, the researchers set up an elaborate deception to keep the experiment spatially realistic. For each trial, the listener was first separated from the scream providers and extras in the chimps’ compartmentalized sleeping room, where it could still see and hear them, and then released into another indoor room where it could only hear the others. After the subject was isolated in the other room for a few minutes, the screamers and extras were released into an outside area where wouldn’t hear their own calls being broadcast.
Since the subject would certainly hear the sound of the hydraulic doors and maybe associate it with release to the outside area, the chimps’ keeper then opened and closed some of the internal doors in the sleeping room and gave shouted commands, play-acting the procedure for moving chimps around in the room and creating the impression for the subject that some unknown chimps were still in there. The researchers then broadcast the call sequence recordings from the sleeping room (here’s a diagram of the chimps’ changing positions through the experiments). The subject’s response to the playback was filmed, and after five minutes the keepers simulated the release of the chimps from the sleeping room to the outdoor enclosure by shouting and operating the doors. The subject then rejoined the group in the outdoor enclosure.
While all this went on, the researchers measured (1) the duration the listener looked towards the sleeping room in the minute before the playback, (2) the duration he or she looked towards the sleeping room in the minute after the victim screams began (that is, where it became apparent whether the scream sequence was congruent or incongruent) and (3) whether the subject approached the sleeping room doors in the minute after playback.
Do you hear what I hear?
Comparing responses to the two scream sequences, the researchers found that eight of the ten chimps looked in the direction of the screams for an average of 3 seconds longer during the incongruent sequence than during the congruent one, one looked longer during the congruent sequence and one showed no discrimination between the two. Additionally, four of the chimps responded to the incongruent sequences by approaching the doors to the sleeping room; three of them did the same in response to the congruent sequence. Below is a chimp-by-chimp breakdown of the looking duration for each sequence.
The differences in looking responses couldn’t be explained away by the acoustic features of the call sequences. The chimps showed a weaker response to the congruent sequences even though these were more acoustically attention-grabbing and contained more call types from more individuals, including a top-ranking male, who generally evokes the most interest when a fight breaks out. Instead, the researchers think that the chimps’ stronger response to the incongruent sequences suggests that the chimps were figuring out the social roles of the two screamers and making sense of the conflict by putting the calls and the roles of the callers in a wider social context. Since the social upset happening in the incongruent interactions couldn’t be sussed out simply by the acoustic features of the call sequence; the listener would have to make some inferences about the direction of aggression by assigning two distinct social roles – victim and aggressor – to the screaming chimps and integrating that with their existing social knowledge about the expected social standing of the screamers. The fact that the chimps seem to have done so suggests 1) that our ability to read into screams, cries and other calls first appeared far back in our lineage and 2) that the gap that separates us from the rest of the animals has narrowed again.
Reference: Slocombe KE, Kaller T, Call J, & Zuberbühler K (2010). Chimpanzees extract social information from agonistic screams. PloS one, 5 (7) PMID: 20644722
Slocombe, K., Townsend, S., & Zuberbühler, K. (2008). Wild chimpanzees (Pan troglodytes schweinfurthii) distinguish between different scream types: evidence from a playback study Animal Cognition, 12 (3), 441-449 DOI: 10.1007/s10071-008-0204-x