Guest Blog

Guest Blog

Commentary invited by editors of Scientific American

Animal emotion: When objectivity fails


There seems to be an explosion of concern over animal welfare these days. With growing awareness to factory farming conditions, Americans are at last faced with the recent histories of their burgers, their nuggets, their pork chops. What we see makes us viscerally uncomfortable, and reasonably so.

Pig_gestation_crateThose of us who are sympathetic to animal issues find it wretched to consider that our dinner might have had emotions. Whatever your beliefs or feelings are toward nonhuman animals and their emotions, it’s not an easy thing to witness, a full-grown pig in a gestation crate, a chicken with only half a beak.

The problem is, we can’t say for certain that nonhuman animals feel anything at all.

In the spirit of full disclosure, I’m not an animal rights activist. I respect what animal welfare activists do, but I eat meat. I wear leather. To contradict myself, I also love animals (far better writers than I, like Jonathan Safran Foer [1] and Hal Herzog [2], have recently explored our weird relationship with animals). Although I don’t identify as an activist, I deeply support the efforts for greater welfare for the animals we "use" for meat, research, and entertainment.

Chicken at Farm SanctuaryThat affinity for animals initially drove me to visit Farm Sanctuary [3], a rescue center for factory farm survivors, late last year in Watkins Glen, New York to do research for my Master’s thesis. I chose to cover nonhuman animal intelligence, both emotional and cognitive, and figured that the farm animal shelter would be a great place to set a scene.

I met lots of chickens, bred for both meat and eggs. When confined to a factory-farm poultry house, the birds are debeaked, which means that about a third-to-half of their beaks (appendages with nerve-endings) are clipped off as chicks. This process prevents them from inflicting injury on other birds, and also discourages cannibalism, a rare act that occurs with higher frequency when the birds are stuffed tightly together.

Cow at Farm SanctuaryAt the Farm, the birds roamed freely, pecking idly and gathering in their little bird-cliques. My host for the day, the visitor program manager Don Walker, explained to me that these birds wouldn’t live very long here, no matter how ideal their feed or space. Chickens have been bred to be giant-breasted echoes of what they once were—flying, perching, foraging birds that actively seek out sunlight and social interaction. Because they grow so quickly, their enormous chests fracture their bones, and when they can no longer walk, they are trampled by the thousands of other birds in their houses.

The stories were the same for the pigs, cows, and goats living on the Farm; all the animals suffered from any number of maladies stemming from past malnutrition and growth-specific breeding. Many, especially the pigs, had scars from stress-induced injury. Stereotypic behavior—characterized by obsessive repetition, sometimes to the point of self-destruction—is common among farm animals confined to tight spaces, and some still wore on their faces and limbs the reminder of their prior living conditions.

I took many notes while I watched the animals go about their business that day. Pigs snorted, grunted and rummaged, occasionally pacing in their large pens. Chickens trotted casually around the yard, bolting and stopping, pecking around, and setting themselves down, tired from the excessive weight they couldn’t shed. Several of the cows I met didn’t stand at all, instead preferred to lie lazily and comfortably in the grass fields.

Goat at Farm SanctuarySo the day went at Farm Sanctuary, watching the meandering residents. My observations inevitably led to a troubling concern: if nonhuman animals experience emotions, then what of these pigs, chickens, and cows? What of all the others, the billions slaughtered in the U.S. each year? [4] Although science looks to find the answer to the perplexing question "do nonhuman animals have emotions," it consistently comes up empty-handed.

The search for nonhuman animal emotion is ongoing and pervasive, regularly making science and political headlines. Even in a recent post here called "Bambi or Bessie: Are Wild Animals Happier?",[5] author Christie Wilcox tackled one angle of the question, though her assumption included nonhuman emotion as at least basically acceptable. And I agree with that assumption, as do many of the most prominent scientists trying to pin down conscious phenomena in other animals.

Chicken at Farm SanctuaryTheir beliefs and feelings toward the matter, however, are intuitive rather than scientific. Subjective rather than objective. This is a big no-no for classical Cartesian reductionism; after all, the scientific method depends on observable, replicable results. In the words of David McFarland, a former Oxford animal behavior professor, "A scientist's hunch is acceptable as a start, provided that it leads to a theory that can be rejected in the face of evidence."[6] It’s the "evidence" part that’s the problem.

Emotion in humans is difficult enough to gauge empirically and often depends on subjective rating systems; you can imagine how much worse the problem gets when an insurmountable language barrier comes into play. A psychologist can ask her human patient outright what he is feeling, and he can tell her about his depression or anxiety. She can then compare her patient’s response with established diagnostic criteria.

Turkey at Farm SanctuaryIn nonhuman animals, science is forced into a world of inference. When tightly confined, pigs will bang their heads on crates, chew on cage bars, and bite other pigs’ tails. Chickens turn to abnormal behavior like cannibalism and feather plucking. Off the farm, dogs respond similarly to humans when taking psychiatric medicine;[7] rats emit high-pitched chirps when tickled;[8] an experiment can show that lab mice exhibit empathy when their cage-mates are subjected to painful injections,[9] judging by basic stress responses and reactions to future pain.

But whether or not these animals are responding with conscious emotions or simply with unconscious and automatic self-survival mechanisms is a mystery to scientists everywhere, no matter what their intuitions tell them. Instead, all that can be done is to interpret abnormal behaviors that coincide with either high levels of stress or pleasure. In this realm, scientists tend to operate in the extremes.

But what if there existed some standardization for identifying nonhuman animal emotion? What if, like with psychiatry’s Diagnostic and Statistic Manual of Mental Disorders (DSM), some general criteria were collected and used in observation for cognitive ethologists who wish to see which animals are feeling, and which may not be? Such a test would have to be based upon human experience, as understanding a radically different concept of emotion is, as of yet, beyond our scientific reach. The test and the criteria might resemble something of an emotional Turing test, with a pig, chicken, or cow, for instance, in place of the machine.

In some scientific cases, more credence seems to be given to machine capabilities than to those of nonhuman animals. For machine behavior, the Turing test has been used for nearly six decades. Developed by British mathematician and computer pioneer Alan Turing, the test arose out of a simple question: can machines think? Turing developed a method to attempt to answer his question.

In its most basic manifestation, a Turing test involves two humans and a computer. Person A sits before a computer screen; person B and the computer sit behind a barrier, out of Person A’s view. Once this is all set up, Person A types questions into a computer, and both Person B and the computer respond. This exchange continues for a while as a casual conversation, and in the end, Person A is left to figure out which of her correspondents is human, and which is computer. If the computer tricks Person A into thinking it’s human, then that computer "passes" the test and is considered potentially intelligent (it’s important to note that the goal isn’t to get a computer to "trick" a human. The goal is to develop a machine that can function similarly to a human brain, at least at a basic level).

Though it has evolved into many variations, each type of Turing test runs on the basic principle of judging artificial intelligence based on human intelligence. In other words, if it communicates intelligently and reasons intelligently, as judged by a human being, then the machine is granted the title of "artificially intelligent." The search for AI has been long, and the Turing test has yet to be successful—which is not to say scientists aren’t still trying.[10]

Cow at Farm SanctuaryWhile there are obvious issues with a direct comparison between a machine Turing test and an animal Turing test (with machines, we’re testing our own creation; animals are a different story altogether), the analogue isn’t without justification. Adapted and applied into cognitive ethology, an emotional Turing test would take a third-person approach to what has long been a first-person problem; it would increase objectivity in an otherwise subjective experience.

The first step in developing such a test would require that a set of basic, standardized criteria be developed; if measures for emotionally or consciously motivated behavioral responses are standardized, then judging behavior against those criteria becomes possible. In other words, if a pig chews needlessly on cage bars and her bodily responses (cortisol levels and skin temperature, as well as other indicators of stress in the mammalian central nervous system) correspond, then that pig can be said to be experiencing anxiety.

Tufts University professor and eminent theorizer of consciousness Daniel Dennett has proposed something similar for use with humans, a practice he refers to as "heterophenomenology." This practice, he argues, is "the sound way to take the first-person account as seriously as it can be taken." In a nutshell, Dennett says that if the researcher both listens to a subject’s inner account of a situation, and then observes the environmental factors, an objective conclusion can be reached about the inner-workings of the subject’s conscious thought processes.

He writes in an article on the subject: "a more constructive approach recognizes the neutrality of heterophenomenology and accepts the challenge of demonstrating, empirically, in its terms, that there are marvels of consciousness that cannot be captured by conservative theories."[11] The conservative theories to which Dennett refers hold back the study of consciousness in general, in humans and nonhumans alike. And though his proposal is meant for human consciousness, the principals could easily be applied to nonhuman animal emotion (Dennett, who is also a cognitive scientist, sometimes turns to the discussion of nonhuman animal consciousness).

Turkeys at Farm SanctuaryIt's a bold proposal, one that many will emphatically argue requires too great a leap of faith. The prospect is arguably rife with anthropomorphism, the supposed bane of all nonhuman animal consciousness studies. But approaches based on sweeping subjective generalizations are common throughout many areas of human psychology. Consider a hypothetical (and oversimplified) therapy session between Jane Doe and Dr. M. A set of criteria for emotion (in disorder form) is already accepted in the psychiatric field—the DSM—and judging by these formerly laid out, generalized criteria, Dr. M can determine by Jane Doe's verbal explanation of her experience whether or not his patient is experiencing anxiety. Dr. M makes a leap to believe his patient, thereby prescribing a medication that will alleviate her suffering. This trust in Jane and how well she knows her own emotion is subjective, but is nevertheless generally accepted. Take the language out of the equation—Jane’s ability to tell Dr. Z about her anxiety—and all we’ve left to go on is inference.

Is it so extreme an idea to suggest that nonhuman animals experience some form of emotion? It’s ethically easier to assume, as did Rene Descartes, that nonhumans are simply mechanized automatons, more like Turing’s machines than like us. I know my visit to Farm Sanctuary would have been much easier if I believed none of the residents had suffered any pain or anxiety.

But I think we (scientists included) know better. The question is, really, can we ever prove it? One Oxford psychology professor, Cecilia Heyes, has hope: "I assume on nonscientific grounds that many animals experience phenomenally conscious states – that they are not ‘beast machines’ - and I find it plausible that, at some time in the future, the presence and character of these states will be discoverable by scientific methods." [12] Without some standards like those outlined above, however, I doubt that Heyes’ future is anywhere near.

The need for objectivity in science is understandable. As a student about to graduate with a degree in science writing, I fully appreciate that the scientific method helps weed out bad science, keeping the wolves out of the proverbial henhouse.

But when inherently subjective issues like emotion are on the table—with deep ethical implications hanging in the balance—I can’t help but feel that objectivity is failing in this case. At some point, it might be that we should, in the words of cognitive ethologist Marc Bekoff, "get over the issue of anthropomorphism and move on—there’s important work to be done." [13]


1) "Eating Animals" by Jonathan Safran Foer

2) "Some We Love, Some We Hate, Some We Eat" by Hal Herzog


4) National Agriculture Statistics Service. "Livestock Slaughter." 21 4 2011. USDA Economics, Statistics, and Market Information System.

5) Wilcox, Christie. "Bambi or Bessie: Are wild animals happier?"

6) McFarland, David. Problems of Animal Behavior. Longman Sc & Tech, 1989.

7) Dogs and psychiatrics: "Reconcile." 18 3 2011. Drug Information Online. 2011.

8) Panksepp, Jaak and Burgdorf, Jeff. ""Laughing" rats and the evolutionary antecedents of human joy?" Physiology and Behavior 79 (2003): 533-547.

9) Langford, Dale J., et al. "Social Modulation of Pain as Evidence for Empathy in Mice." Science 312.5782 (2006): 1967-1970.

10) See the Loebner Prize Competition for more details on the ongoing pursuit for Turing’s AI.

11) Dennett, Daniel. "Who's on First? Heterophenomenology Explained." Journal of Consciousness Studies 10 (2003).

12) Heyes, Cecilia. "Beast Machines? Questions of Animal Consciousness." Frontiers of Consciousness. Ed. M Davies and L Weiskrantz. Oxford: Oxford University Press, 2006.

13) Bekoff, Marc. The Animal Manifesto. New World Library, 2010. p. 76.

Image credit: Kristina Bjoran

About the Author: Kristina Bjoran is a student in MIT's Graduate Program in Science Writing, where she writes about technology, environmental studies, and nonhuman animal intelligence. During her down time, she volunteers with animal shelters, writes for nonprofits, and dabbles in photography and skydiving. Follow Kristina on Facebook and Twitter, and visit her page on MIT Scope.

The views expressed are those of the author and are not necessarily those of Scientific American.

Please take a minute to vote for Scientific American in the Webbys: We are a nominee for excellence in science.

The views expressed are those of the author and are not necessarily those of Scientific American.

Share this Article:


You must sign in or register as a member to submit a comment.

Starting Thanksgiving

Enter code: HOLIDAY 2015
at checkout

Get 20% off now! >


Email this Article