What is the best way to determine the function and context of animal calls?

Question

I've come across multiple bioacoustics studies that attempt to describe the function of a particular call made by animals - i.e. by ascribing meaning to stereotyped calls that are given in particular contexts.

Some examples include;

• echolocation clicks used for finding food in bats and odontocetes, or;
• "loud/long" calls of many primates, such as gibbons or howler monkeys, that are associated with pair-bonding or territorial defence, or;
• sexual advertisement calls of frogs, or;
• contact calls given by ungulates for social cohesion in groups.

In a lot of other instances there seems to be no way of knowing. For example, it seems like most aquatic species have unknown functions for their calls, or that there is extreme difficulty in understanding those functions.

I have three questions:

1. What are some of the more straightforward ways to frame hypotheses that test the functions of these calls?

2. How can we be sure that what we are contextualising is actually related to the physical properties of those calls?

3. Are there any decision trees or resources for determining the functions of calls? For instance, if a type of call (e.g. echolocation) is given in a specific context will that call always be considered "functionally specific", regardless of whether the animal is a human or a bat or a dolphin?

Answer 1

Great question! I will only propose an element of response to this question, there's probably much more to it: I think behavioural observation would be the first step to build hypotheses about the function of a certain sound. Then, for signals or cues, I would think the most helpful method to establish a causal relationship between the sound and the behavioural response is to use playbacks and observe the behavioural responses of the receivers. This will not be possible for all animals and all signals but I think this applies to many signals and is the best way to establish function. This wouldn't work for sounds that are not signals such as echolochation clicks though.

Answer 2

The best way is to correlate the sound with visually observed behaviour. Obviously number of observation must be such that cause and effect can be deduced and not only correlation.

For me an example could be: alarm call of some terrestrial species

you hear the call, but if you do not see the predator then you only can speculate that "this type of call is known to be emitted in presence of a predator, so you conclude that also here a predator is present and this call constitutes an alarm call". Such a conclusion is only valid if the species is well studied and no counter examples are known for the functionality of this particular call.

Echolocation clicks of, say sperm whales seem to be easier to interpret, but they are not. Is the function to find individual prey items or are they emitted to get an overview of the prey-field. They are technically suited for both functions, but without understanding of the cognitive capabilities it is hard to assign a primary function to them. And there will always be a third minor function to them (communicate to conspecifics and other un-heard functionality :-)

IOW, the function of calls and songs needs a detailed understanding of behaviour and capability, which with acoustics alone is difficult to establish.

Answer 3

I agree with the previous comments! And just to throw more uncertainty into it, as a primatologist I can say that the field still debates the function of primate "loud calls" - why howler monkeys howl is still very much up for debate, for example.

I study ruffed lemur vocalizations, and we've matched up vocalizations with behavioral contexts to infer functionality/info being transmitted as the previous commenters talk about. This includes information on the behaviors exhibited right before and right after calls, demographic information about the caller(s) (they produce contagiously chorus'd calls as well as individually given ones) and the audience (e.g., number of callers, number of individuals within 'x' distance, sex/age class of callers and/or audience), changes in group size or composition before or after calls (ruffed lemurs exhibit fission-fusion dynamics so this changes frequently), calls given from out-of-sight groups, and other potential influential events (e.g., a group of sifakas [different lemur species] came through, a plane flew overhead, etc.)

From these inferences, we then do playback studies to further finetune our predictions. Playing back different conspecific call types to groups of varying characteristics and observing responses. And also playbacks of e.g., a hawk shriek to assess vocal response to predators.

And just to confirm the uncertainty of this guessing game, we found very inconclusive evidence that the ruffed lemur "long call," the roar-shriek (youtube it if you really want to be woken up) aligns with any particular behavioral context, audience, etc. And we continue on... :)