The ironic thing is of course that I do not have enough time for this blog, but then ... it is a challenge ... :-)
This will focus on making machines more emotional or deal in more intelligent ways with human emotions. Check it out ...
I decided to start a second blog that focuses on Affective Computing. It is called Affective Computing Science and it will be a bit more technical and also contain links and or comments in that special field.
The ironic thing is of course that I do not have enough time for this blog, but then ... it is a challenge ... :-)
This will focus on making machines more emotional or deal in more intelligent ways with human emotions. Check it out ...
2011 – the 20th anniversary of Nirvana’s NEVERMIND, Pearl Jam’s TEN, and Fridlund’s paper on IMPLICIT SOCIALITY. While I have seen whole special issues and disks covering Nirvana’s album and some long articles on Pearl Jam’s first album, I have not seen any special issue, or even long article on Alan Fridlund’s article. A shame. A lacuna that needs to be taken care of, so let me remind you of, or introduce you to:
Fridlund, A. J. (1991). The sociality of solitary smiles: Effects of an implicit audience. Journal of Personality and Social Psychology, 60, 229-240.
This is a very important empirical study that has serious implications for the way that we (should) think about facial activity. Everybody who works in research and engineering on projects that involve the relationship between what people feel and what people show on their face should know about this study and its findings. They should know it well.
Imagine, you hear about a psychological study that involves a task you should perform with a friend – so you bring a friend along to the lab. Once there, the experimenter tells you that the two of you are going to watch a funny movie together (actually a series of five clips, spliced together, totaling a bit over 13 minutes). While watching the clip, the activity of a couple of facial muscles is measured using electrodes glued to your face. If you must know, a pair of electrodes on your left cheek, your right cheek, and the left brow. The cheek site relates to the muscle Zygomaticus Major, that we use for smiling. The brow site assesses activity at the site of Corrugator Supercilii, a muscle we use to frown by pulling the brows together and down. To make it less explicit that facial activity was at the center of the researcher’s interest there were also some electrodes on other sites, the top of the head or the hand and the experimenter said something about recording brain waves … After the end of the funny clips you answer how you feel using some scales, in fact you had also answered the same questions at the beginning of the experiment. Not complicated. (Fridlund refers to this condition as explicit coviewing)
In a parallel universe, you arrive with your friend, but the experimenter in this world tells you that you will watch a funny movie in one room while your friend watches the same movie at the same time in a different room. Again, your facial activity is measured and you indicate after the experiment how you feel. (= implicit coviewing)
In yet another universe, you arrive with a friend, but this time the experimenter tells you the good news that you are going to watch the funny clip while your friend is filling out something that does not relate to this at all – it is part of a different study. (= implicit irrelevant task)
Surprisingly, there is even a parallel universe where you are not aware of any experiment involving friends – you come to a study alone where you are being told that you will watch a funny movie etc etc. (= solitary participation)
So now that you picture the four situations in your mind – like in Sesame Street: One of the foursituations when you are watching the clip is different. Which one? The one where two people watch a movie together – in the other three you are alone in a room watching the movie, having electrodes attached to your face, and answering questions.
According to the predominant view regarding the relationship of feelings and emotional expressions at the time, one would predict that there should be no difference in the expression across the three solitary conditions – or, for that matter, the subjective experience in the three conditions where you are alone. Why? The assumption would be that when we are alone, we show simply what we feel – which would be the same in all three. There is no need to modulate what you show for reasons that it might be inappropriate, or that somebody would expect something from us – in contrast, when we are together with others, then different rules would apply – what Ekman and Friesen have referred to as “display rules”.
I remember, when I was in grad school at Dartmouth College, in the late 80’s, I received a description of this study and was asked what my predictions were for the findings. I do not remember exactly what I answered, but I do remember that I was wrong, once the results of the study were revealed – The ingenuous study by Alan Fridlund demonstrated that the most smiling occurred when friends watched the clip together, OK, no surprise here, but rather than all the other situations being the same, it mattered what the subjects thought their friend was doing. If you thought your friend was watching the same film in another room concurrently, then you smiled more than when your friend was supposedly doing something else. For Fridlund this was not surprising – he argued that it does not only matter who is actually physically present, but who is present in our head – something he referred to as implicit sociality. The next surprising thing was that the subjective experience was the same all cases independent from the intensity of the smiles – from this Fridlund concluded that emotions do not matter – that facial activity is determined only by the social context – whether explicit or implicit.
I do not agree with the latter conclusion – and I have reasons not to, but I will talk about that in the next installment of the blog. In the meanwhile, if you find this interesting, go and get a copy of the paper and read it – including the theoretical explanation that relates to behavioral ecology – Fridlund explained this much better than I could do here in a couple of paragraphs. In fact, you will find that there were also a couple of hypothetical studies – one involving “a small sample (n = 5) of advanced graduate students and doctoral candidates studying the face and emotion at a major northeastern university” (p. 235) that did not predict these results (to illustrate the point that the pattern of findings is not trivial) … you know one of these students I guess ;-)
I think I have seen more robots in the two weeks of my Japan visit than ever before. Obviously, there is much research being done on all aspects of creating robots/androids and studying the interaction of humans with artificial entities, whether they are physical, as in a robot that moves and might be able to move around, or on a computer monitor as a graphic representation.
One of the highlights of the ISRE conference was a key note address by Hiroshi Ishiguro from University of Osaka and the ATR. Prof. Ishiguro is well known for his work on creating machines that look like humans. Specifically, like particular humans - hence he refers to these as geminoids - as in twin. A different term is "actroid" - geminoid is in fact a registered trademark by ATR. Whether geminoid or actroid, they are all androids (or gynoids), i.e., machines that resemble male or female humans. The effect is stunning. Whether moving or not they are close enough to a "real" human being to have some sort of presence. Unless they are sitting in a corner and not doing much, they are not mistaken for being human (von der Pütten et al., 2011). However, they are also not "just" a machine. How much of humanity we perceive in these machines is one of the key research questions several research groups are currently investigating.
Von der Pütten, A.M., Krämer, N.C., Becker-Asano, C., & Ishiguro, H. (2011). An android in the field. HRI 2011, 283-284.
Hiroshi Ishiguro, Geminoid F, Arvid Kappas at ISRE 2011 in Kyoto, July 2011.
Ishiguro outlines the history of his quest for building machines that we might perceive as being human. One of the early attempts was a gynoid that looked like his daughter (Repliee R1). In his words - that was uncanny. But he believes that the newest generation of geminoids, such as Geminoid F is not uncanny anymore. The word uncanny here refers to Masahiro Mori's notion of the uncanny valley. Personally, I am not sure regarding Geminoid F. To me she was still somewhat uncanny ...
The official conference fan. Not a bad idea, given the weather.
I am currently at the conference of the International Society for Research on Emotions which is being held in Kyoto, Japan, from July 26-29, 2011.
So what can one find at such a conference? Interestingly enough, not only the present, but much past, and also some future. Today, the first day of the conference, there were some presentations dealing with how emotions, such as fear or anger, were perceived by philosophers in Greece and Rome (David Konstan, Brown University) or in England between the 12th and 18th century (Susan Broomhall, presenting for Philippa Maddern, both Australian Research Council Centre of Excellence for the History of Emotions). What can one learn from this? That the way we understand what emotions are, how they come about, what they do etc. is shaped in powerful ways by the environment and the prevalent beliefs and practices. This is relevant because it is unlikely that today we are so enlightened that we have achieved objectivity in answering these questions. In fact, not only is objectivity a difficult concept today and in the future, when emotions are concerned, but the questions that we ask, are also shaped by our present cultures, as two other presenters pointed out. This matters because, as David Konstan pointed out, if we do brain research, on a particular emotion, say anger, then how we go about doing this is strongly shaped by our present understanding of what anger is or is not - thus, the objectivity that the machines and sophisticated methods in neuroscience seems to provide is hopelessly intermingled with the subjectivity of the researcher that defines the question to be studied there and how. Interesting.
Tomorrow there is much about the future, robots, cyberemotions, and the like.
What an excitement! Soccer is grabbing the world. Possibly more so than ever. South Africa is a gracious host to the 2010 FIFA World Cup, as it is officially known. Surely an event that is accompanied by many emotions - on the field, in the stadium, in the press booths, GOOOOOOAAAAAAAAALLLLLLLLL!, at home, and wherever people watch and listen to the matches as they happen.
Emotions are not just intrapersonal events - processes that happen within us, but they happen between us. Just how social emotions are is one of the areas of debate. For example, some researchers will acknowledge that emotions are social because many of the events that trigger emotions are social. For example you find out someone lied to you, you meet your loved one after a long separation, someone cuts in and takes "your" parking space. All well. These are social situations, and the importance of these situations is high because of the implications that others have for us. There will be hardly disagreement on this one. In fact, the same experience can evoke different reactions if you believed it to be caused by a software program, rather than a person.
Take the following, by now classical, example: One of the experimental paradigms that has found some interest in a new field called neuroeconomics is the ultimatum game. Imagine that you come to my experiment and I put you in a room with someone else. I give you 10 Euro and tell you that you have to split that money with your fellow subject any way you want it. If she agrees to the split, both will get the money. If she does not agree then neither will get it. Fairly straightforward. You offer 5 Euro, she accepts, and both walk with 5 Euro. But what happens if you offer only one Euro? Most likely, studies have shown, the other will reject this offer. This phenomenon caused (surprisingly) some head scratching in economists who had a hard time understanding why that would happen. Why would you reject money with no strings attached? Well, to many people outside of academia it appears quite logical that unfair behavior can tick you off and that one way to punish the person treating you unfairly is to make sure that the other one will not benefit from that unfair behavior. So far so good. Now, one of the interesting studies in this field by Alan Sanfey and colleagues (2003) showed that if the offer is made by a computer program, rather than a person, the unfair offer is more likely to be accepted. This is important because it demonstrates this facet of the social nature of emotion - the same thing happens to you, but in one case you believe there is a person causing this, in the other you believe that a computer is causing it and your brain and the rest of your body behave differently. An elegant way to demonstrate the importance of social factors for the causation of emotion.
However, I am a firm believer that emotions are social in many more ways. One aspect that an event like the world cup highlights for us is the complex forth and back that happens in mass events like a football match. Emotions are often caused by social events, they have social consequences, but in addition, in the context of communication processes there are complex real-time things happening that apparently have emergent properties. There are very complex networks at play - you might watch with some friends, you might watch with people you do not know, but who, by their fan clothing and their behavior betray that they are "with you", there are national groups involved, there are teams involved, all of these form invisible ties that make you respond to their behavior, and those who can see and hear you to your behavior. Sometimes it is even sufficient to imagine the others (we'll talk more about this one several times). It is a real time network of subnetworks that involve feedback loops that lead to big effects: Is this the madness of the crowd that LeBon wrote about? I believe it is more complicated. In the context of our CYBEREMOTIONS project we are actually interested in such phenomena on the Internet (more about that one also).
So I leave you with the challenge to think about the complexity of the communication and feedback processes in real time while watching the next world cup matches. If you can, go to a public viewing area, and observe what is going on as people stare at strangers on huge displays, staring at other people in a stadium, staring at themselves on displays in the stadium, or staring at the people on the field, who are emoting, a lot! Enjoy the emergent chaos, and try not to get too annoyed by the vuvuzelas!
Sanfey, A.G., Rilling, J.K., Aronson, J.A., Nystrom, L.E., & Cohen, J.D. (2003) The neural basis of economic decision-making in the ultimatum game. Science, 300, 1755–1758.
Thank you Elena for the question in the comment to the last post - could it be that emotion is a sort of sense - like the 7th sense?
Of course much depends on the definition of what a sense is. The five senses that are usually referred to, sight, hearing, touch, smell and taste have specific sensing organs - receptors that are to a certain degree specific to some stimulation, for example chemical, or pressure. In fact there are more senses, such as nociception, or thermoception. In addition, there are senses that deal with stimulation inside of the body. Emotions in contrast seem to be aroused by the results of other senses, something we see, or hear - or imagine seeing or hearing. Thus, there seems to be a big difference. Emotional processes (let's not go back to the definition issue for the time being) follow other sense data, are not dependent to a specific modality, and, we now know, are linked to relatively complex networks in the brain that are not modality specific.
But is intuition not "the sixth sense"? Well technically not. More often sixth sense is associated with the notion of paranormal phenomena, such as telepathy. Independent of the fact that there is no scientifically validated evidence of such paranormal phenomena, the use of the term sense in this context is colloquial.
But what could it then be? I am at this point wondering whether augmented reality is a useful metaphor for the relationship of emotions and perception. Emotion can provide a different dimension to sense data - the dark ally that feels creepy, the cake that looks yummy, the dog that looks dangerous. In all of these examples words are used to convey a feeling that is the summary function of a rapid and typically automatic evaluation of properties of the object(s) perceived in relationship to their implication. One aspect of emotions is that they send nonverbal messages that are shorthand for the implication. They may trigger in human verbal associations and interpretations, but they need not. Important: these messages need not be consciously received. After all, part of the process is the rearrangement of resources to deal with situations, and that is the case whether one is aware of it or not.
Wait a moment ...? Unconscious emotions? Yes, If you did not already do so, please check out
Winkielman, P., Berridge, K. C., & Wilbarger, J. L. (2005). Unconscious affective reactions to masked happy versus angry faces influence consumption behavior and judgments of value. Personality and Social Psychology Bulletin, 1, 121-135.
Thanks so much for the feedback on the list ... I have received some comments to my previous posting “pain and pleasure” via the blog, but also via email, and via facebook. All of these are of course welcome and I will use them to make more lists and entries real!soon!now! I appreciate the encouragement, and also the challenges ;-)
Right now, I am still in the process of thinking about what to write in the next weeks and months. For me, this blogging thing is new - it is interesting, exciting, daunting, and also, at times, perplexing. Yet, curiously, problems in planning the blog resonate with problems in conducting emotion research. Topics here are: levels, dynamic aspects, intensity, and function.
LEVELS :: In the emotion context, levels may refer to levels of analysis and specifically the approach put forward by Cacioppo and Berntson (1992) when introducing the notion of Social Neuroscience. I love this idea very much; it had a huge impact on my thinking, my teaching, and also my research. The idea here is that some researchers might focus on something like the immune system, others on psychological processes, such as stress, yet others, on interindividual differences, and someone else on the role of social support networks for well-being. Typically, in each case the focus is on a single level of analysis, but the mantra Cacioppo and his colleagues have been repeating ever since is that to understand the type of phenomena that we are dealing with, one a) needs to take into account how the same process relates to different methods, theories, and observations at different levels and b) how these levels relate to each other and interact. In the example, knowing about the inner wokings of the immune system will help in understanding how psychological processes at the individual level, and here interindividual differences, relate to social support. In a nutshell. I will get back to this issue of the necessity of multi-level approaches often. If you do not know the classic paper by Cacioppo and Berntson, the reference is below. John Cacioppo's page lists many relevant papers and gives also access to many. An excellent place to check out!
For the blog context, levels refers to how I should discuss the issues I want to discuss – for the specialists (and I know that there are some who are checking out this blog), or for the colleague with some interest in emotion, for students, or for somebody who is not interested in the nitty gritty of academic discourse, but who wants to understand something about emotions and wonders how such a complex and subjective thing could be studied by science? Frankly, I have not fully decided what level should be the target level. It is sure, that I will sometimes try to be more on the introductory level, and at others, go into some of the more arcane details of specific theories – for sure, there are parts of my website arvidkappas.com that simply list some starting points for the non-specialist – such as, what to read, or where to go to find out more. However, as a rule, the topics here, and the language chosen, will be a bit more on the advanced side, assuming the reader has already some familiarity with key theories of emotion. Feedback is most welcome. I am aware, that it will not be possible to do everything for everybody. I will also try to make cross-links to other blogs so that the interested reader might serendipitously find something even more interesting than this blog ;-)
DYNAMIC ASPECTS :: Dynamic aspects relate to how things develop over time – things like duration, fluctuations. This is a major problem in emotion research. We know so little about how long an emotion lasts, how one emotion tends to change into something else, how frequent we smile, how often we show certain physiological responses. One of my own interests relates to the role of dynamics in facial expressions, particularly smiles. How fast does a smile have to show up on the face for you to feel that it is genuine and not fake?
With regard to the blog it means, how often can and should I write? How much time should I leave between? Should I try to have a fixed rhythm? I’ll see.
INTENSITY :: How intense should this get – meaning, should I provoke? Should I challenge specific readers? I intend to be relatively spontaneous, as I am when I teach – this is not a scientific article, it is a soap box, so I will see whether this works or whether I am burning my fingers as I go along. Maybe I might get a bit out of control? That would be fitting for a blog on emotions, I guess.
In the emotion world, intensity is a tricky one – sometimes it feels as if, for example, anger and slight irritation are the same thing, but at different intensities, and indeed researchers have made that point. Others believe these are different states with different properties Also concepts such as moods or temperament are not only linked to temporal differences, but also to intensity differences. How to measure the intensity of an emotion is no doubt linked to how I define emotion – so is the definition issue, alluded to in the previous post, really a red herring, or is it more important than that?
FUNCTION :: I believe strongly that a functional analysis of emotions is key to understanding them. Emotions are not there to amuse or to puzzle us – they do things. We have evolved emotions because they are particularly good in dealing with certain types of situations. Sometimes they are not so useful, but overall we are much better off with than without. This is a notion that goes back to Darwin’s analysis, particularly of emotional expressions, but also to all behaviors and biological systems. A functional analysis can also be misleading, but by and large it’s the best we can do. What that means, I will say in more detail later, as I will pick up on these other topics. I apologize to Tom – your comment was well taken, much abstract, little concrete in these opening posts …
But what is the function of the blog? Should the function(s) and goal(s) not shape the form and the content of what is being discussed? So I should know what this is about before I start. In an ideal world this might be the case, but in reality, well, this is difficult to achieve. Surely, there is more than one function I guess. Some relates to vanity and hubris. Some relates to idealism and naiveté. Some relates to me trying to make meaning out of what I can observe. This however, for me, is a distributed task – it is about making meaning for myself, but also being part of a larger enterprise that makes meaning out of observations linking many brains - this is why we teach, why we have schools and universities. Emotions are such wondrous things, it takes many brains and much time to discover some of the whys and the hows. The internet is such a curious thing, it allows connections in unexpected and rather direct ways, so I guess what it really is, is an experiment that is trageted at joint discoveries. However, I will also try to entertain a bit, along the way, and I hope I succeed.
So for now I ask you to bear with me until I have a better grasp on the blog, relating to LEVELS, DYNAMIC ASPECTS, INTENSITY, and FUNCTIONS. So that I can talk about LEVELS, DYNAMIC ASPECTS, INTENSITY, and FUNCTIONS of emotions.
Cacioppo, J. T., & Berntson, G. G. (1992). Social psychological contributions to the decade of the brain: Doctrine of multilevel analysis. American Psychologist, 47, 1019-1028. PDF
So of course my immune system reacts to the fact that I have reached the critical end-of-semester deadline and under much stress submitted all grades on time. It decided squarely that once that had been taken care of, it was time to multiply all those happy little thingies in my blood stream and do some about this and that thing that has been bothersome to my immune system all along. Three days down the road, two teeth less, and, on the upside, a good level of drugs in my blood I am sufficiently delirious to try to identify some of the big topics that have been taking up time, paper, and energy in emotion research over the last century or so. Of course, I surely have forgotten important issues when preparing this list ad-hoc and I invite comments from readers. I will use this top 10 list to pick what to write about in the coming weeks, but of course not exclusively. The issues are numbered from 1 to 10, but the sequence does not imply anything with regard to importance or hierarchy.
I apologize for those readers who start to be interested in emotion science - the points mentioned on this list are often too brief to understand what issues I am referring to - this is more something for those already in the know - but as I said, I will discuss these things in the coming weeks. I will then also try to outline what the problem is, and as usual, try to put some useful references and links in there. BTW - I am open to wishes as regards which topics I should attack with preference.
The release of the Rolling Stones' Exile on Main Street was the excuse I needed to finally plunk down some electronic cash and get the Rolling Stones Boxed Set. Among the treasure trove of all things Jagger/Richards also Emotional Rescue - by coincidence - or is it - the name of an article about some of the work my colleague Heather Urry is doing at Tufts University.
The research described relates to regulating emotion - apparently the topic of the week for this blog. The article in Tufts Journal is worthwhile reading and might be a good starting point to find out more about Heather's work. What I find fascinating is that in my mind much of the work on the cognitive regulation of emotion goes back to the pioneering work of Richard Lazarus fifty years ago. This is not a criticism regarding today's work but a praise to this early groundbreaking work. Already then, Lazarus and his colleagues demonstrated that the physiological response to a stressful film could be manipulated by either a text presented while the film was shown or given before to create an orientation that influenced the meaning of the bloody film. What these studies showed is that there is no fixed link between a stimulus and the emotion it would elicit, but that how the stimulus is appraised - what the meaning of this stimulus is - plays an important role. This was a revolutionary statement then and to some degree it still is now. It is the basis of teaching people to control their emotions by trying to think about certain things in a different way. GIven how much modern emotion research focuses on the brain it also underlines the plasticity of emotional processes. The brain might have a particular sensitivity to some things, such as faces, or snakes, but in many cases the reaction to a particular pattern is not fixed. The intermediate step between a stimulus and a response is the meaning of that stimulus. Today, rather than just looking at peripheral measures of emotional arousal, such as skin resistance, we can observe how brain activity in specific locations, such as the amygdala, is affected by how we think about a stimulus, such as an ugly picture. Consider the study by McRea and colleagues below as an example.
One of the most interesting challenges in emotion research is to integrate two types of processes - those which we can control by thought and automatic processes that often even happen outside of our awareness. I will be talking more about that ...
Now back to listening to some Stones ...
McRae, K., Ochsner, K. N., Mauss, I. B., Gabrieli, J. D. E. & Gross, J. J. (2008). Gender Differences in Emotion Regulation: An fMRI study of Cognitive Reappraisal. Group Processes and Intergroup Relations, 11, 145-162.
I remember very well that in the early stages of my PhD thesis I had intense discussions with John Lanzetta regarding emotion theory. John was a very emotional emotion researcher! Content-wise, I had been very much influenced by the second wave of appraisal theories that appeared at the time. Coming from Klaus Scherer's lab - in Giessen, Germany, at the time -I was very much into his component process model. At about the same time Craig Smith and Ira Roseman were publishing their seminal papers about their work with Phoebe Ellsworth.
Eventually John asked me to write an overview over emotion theories. I do not remember how long it took me, but the way I remember it, it was a massive file, amounting to about 100 pages. Probably it was less though. In any case, once I was done, John read it, approved of it, and said that I should put it away now. He said, there was no need to put it into the thesis itself because nobody would want to read it, but instead focus on the research I was going to do. I was quite furious. What a waste, I thought. Of course, that was not true. I have been talking about the history of emotion research ever since - I would call that a sleeper effect ...
However, the important lesson here was clearly not to get bogged down in some of the endless discussions between theorists. Know the theories yes, but do not waste too much time picking a fight. John pushed for data, empirical observations, and models that could make sense of these data. There was no question in my mind after discussing with him that emotions were foremost social processes. Thus, the regulation of emotion would always involve a social component. We regulate emotions in part because of social context, just as we regulate our social context with emotions (for example emotional displays). Our emotions are regulated because of the way that the different emotional components are connected in feedback loops. If the social rule states not to smile, and not smiling moderates how you feel, then social rules can modulate emotions via expression. Similarly, empathy is one of the bonds between people that imply that emotions are not just something inside of us, but also something between us. Emotions are social!
Today, when discussing these things I like to use expressions like
Emotions are self-regulating processes that serve nested intra- and interindividual regulatory functions. Psychological and neuroscience theories that deal with emotion regulation as an after-thought are bound to fail capturing the complexity of multi-level regulation that is part of typical emotional episodes.
Kappas, A. (2010). Emotion/Regulation: Never Tear Us Apart. Presented at the Emotion Preconference 2010 to the Society for Personality and Social Psychology Annual Meeting. Las Vegas, Nevada, January 28.