Psychology is relevant to the humanities.

Posted Dec 15, 2013

text book

and the text was with God,

and the text was God.

~The Book of Text

Au commencement, il y'avait le mot juste.

~ The Book of Made-Up Quotes

In 2011, I went to Rabat, Morocco, to participate in the VIth International Congress on Humboldt. For me, this was not the usual trip. It was my first time in Africa and in a Muslim country, and it was my first encounter with the humanities, with men and women of letters. During the three short days in Morocco, I experienced two culture shocks. The first was my exposure to everyday Moroccan life with all its unfamiliar sights, smells, and sounds. It was a positive culture shock, as far as culture shocks go, perhaps because of some unacknowledged stereotypes about the Moroccan people. I came away with  good impressions and pleasant memories. The second culture shock was my exposure to humanist conferencing, an experience that I am still metabolizing today.

I knew, or thought I knew, that humanists are concerned with language, or more specifically, with text. To the few humanists that I know personally, virtually everything seems to reduce (or enhance) to text. I am putting the word “text” in quotation marks because it is ideologically loaded and self-referential. I also presume we are all thinking of written text and not of spoken words. In Rabat, I learned that the humanists’ love of text works itself out in their reading their presentations. I was dumbfounded by the experience. My own field is experimental psychology, and our industry standard (or professional norm, if you will) is to present our ideas simultaneously as visual images and as conversational speech. We draw up powerpoint presentations and tell stories about the images without using notes. In other words, experimental psychologists and other empirical scientists have returned (or regressed, if you wish to be Freudian about it) to an earlier period of human communication, the era of the oral society. An era, which predates the myth of origin as revealed in The Book of Text.

I found the humanist convention of literally reading text off the page, text that they themselves had written, shocking. The shock was not eased by the fact that most presentations were in French, Spanish, or Arabic. I confess that the papers delivered in French had a certain fascination to me. Many of these papers were recited with great drama and ceremony, as if the speaker was trying to evoke a higher power in order to bring the text to life. My favorite presentation was by Professor Nitschack, who is a German living and teaching in Chile, teaching Brazilian Studies, and who declaimed his presentation en bon francais. He spoke about the légion étrangère. How apt that was.

In Rabat, I made the argument that experimental cognitive psychology has something to say about text. I tried to make my point by way of example (Krueger, 2013). The example was Philipp von Hutten, a German imperial knight and conquistador, who knocked about Venezuela for 12 years until a certain Juan de Carvajal, who had usurped the governorship, and who himself was later tried and executed by the Spanish Court of the Indies (Schmitt & Simmer, 1999), ambushed and killed him in 1546. To me, the psychological question was ‘Why did von Hutten remain in Venezuela after his first and failed entrada?’ Hutten knew that the Colombian gold had already been plundered by a Spanish expedition, he was in debt, and his family had offered to bail him out. So why did he not sail home? Why did he muster a second expedition, went deeper into Venezuela and into debt, and was ultimately killed by other Europeans, a fate that he himself had foreseen as a possibility.

The text to which I had access was a volume with von Hutten’s edited papers, comprising a travelogue and various letters to his family. The volume was compiled by a historian, Professor Eberhard Schmitt, and a descendant of the Hutten family, Sir (Freiherr) Friedrich Karl von Hutten (Schmitt & von Hutten, 1996). The editors provided a careful introduction to the material, but as historical empiricists, they largely refrained from interpretation and speculation. Yet, the facts rarely speak for themselves – however much we would like them to. We seek meaning, and we attempt to construct it through interpretation. The two main pillars of support in the quest for meaning are other historical facts provided by the context and the tools provided by academic theories and methods. For me, these tools come from experimental psychology in general and research on judgment and decision-making in particular. Remember that my main concern was to understand Hutten’s decision not to return to Europe, and to press on instead with a venture he knew had little chance of success.

I felt confident that I could isolate two psychological elements from the text. The first was the concept of honor. Psychologists and anthropologists note that even today many societies are characterized by an ethic of honor (e.g., Sicily, Kazakhstan, parts of Latin America, and the North American South; Gilmore, 1990). The notion of honor comprises a broad range of norms, expectations, motivations, and prescriptions for behavior (Leung & Cohen, 2011). If fully internalized, as in Hutten’s case, a code of honor does not leave much room for choice. The honor-bound person must act – often in extreme and highly risky ways – or else lose face, which may seem like a fate as bad as – or worse than – death. Hutten was clear in his letters that he would not tolerate returning to Germany as a failure, that is, without honor. Today, for those of us who do not live in a culture of honor, it is tempting to dismiss such sentiments and regard them as incomprehensible or irrational. We might think that von Hutten acted irrationally in the sense that he went against his own physical and material self-interest.

But there is another possibility. If a culture of honor is a bundle of social norms, it is not only personal but also social in nature. Indeed, social norms are – by definition – social. As Cristina Bicchieri (2006) explained, norms are the grammar of society. The behavior of Hutten and other honor seekers supports the broader society, which is regulated by the honor code. By staying true to the demands placed on an imperial knight, Hutten reinforced the idea of what it meant to be an imperial knight. From this point of view, it is perhaps even more surprising that his family insisted that he come home. His family placed greater value on personal and family welfare than on the cultural code.

The finding that individuals sometimes subordinate their own physical and material interests to the demands of social norms is a general one. Today, in societies with market economies, we find a strong ethic of fairness and egalitarianism. In studies on behavioral economics, we see that individuals, even if they themselves are not involved in a transaction, are willing to use some of their own resources to punish those who have defected against others, for example, individuals who have not contributed to produce a public good (Fehr & Fischbacher, 2004). These punishers make a personal (i.e., seemingly irrational) sacrifice to sustain a social norm. Arguably, then, Hutten made a contribution to the system of norms and the moral code within which he was brought up. Tragically, he did not realize that the knightly code of honor was disintegrating during the Renaissance. As a conquistador in South America, Hutten was at the same time a man of the past and a man of the future.

The second psychological concept I tried to bring to bear was the notion of sunk costs. A rational person makes decisions with regard to the future (Arkes & Blumer, 1985). He seeks to optimize the consequences of a decision. The past is irrelevant because it is lost. When people honor sunk costs, they make bad investments in the future because of bad investments in the past. They throw good money after bad, and when at war, they are prone to accept more fatalities as if this could somehow justify the deaths that already occurred. An everyday example of the sunk cost fallacy is when you stay with a bad date for the rest of the night or in a bad marriage for the rest of your days simply because you think that a timely disengagement would be a betrayal of your past decision. Notice that the sunk-cost fallacy is also based on a mistaken respect for a social norm. Here, the norm is Thou shalt act consistently. Hutten could have and should have cut his losses when he had the chance. He remained true to his mission when the rational basis of that mission had evaporated, and he paid with his life.

In the laboratory you demonstrate the sunk-cost fallacy by giving respondents descriptions of scenarios that leave no doubt that the consequences of continued action are worse than the consequences of aborted action. If respondents go ahead with the action anyway, because they feel committed to their past choices, they provide evidence for the sunk-cost fallacy. About 30% of respondents in these experiments honor sunk costs. From an experimental point of view, this type of study permits a forward predictive inference. When an unambiguous sunk-cost scenario is set up, we can infer that the probability of a sunk-cost error is about .3. In the analysis of behavior, and in the interpretation of text, however, we are interested in the reverse or causal inference (Krueger, 2014). Once we have observed a particular behavior, we want to know what caused it. In a sunk-cost experiment, the reverse inference is easy, and it is stronger than the forward inference. Once we see a respondent choose a critical behavior or investment, we know that it is an instance of the sunk-cost fallacy because all other causes have been ruled out by careful scenario construction (see Cox & Wermuth, 2001, for a primer on statistical models of causation).

This is an unusual asymmetry. Typically, reverse inferences are harder than forward inferences. When we don’t have the luxury of a controlled experimental environment, when we try to infer causes in the messy world outside of the laboratory, multiple potential causes compete for our attention. Why did von Hutten sail to America? Because he wanted gold and fame? Because he wanted to impress his bride? Because wanted to get away from his family? Because he was beguiled by the stories of Cortés and Pizarro? Because the earth is round? When we look for Aristotle’s efficient cause, the availability of multiple causes is frustrating. The more causes there are, the less significant each individual cause will be.

Spiting Aristotle, we may insist that all causes matter and construct a narrative that combines all of them in a coherent pattern, in which each individual cause appears to be necessary. The construction of such a unified narrative is tempting. The more detail we add to a story, the easier it is for us to imagine it. The more the various parts are interconnected in the story, the more plausible the whole appears to be. For example, to say that Hutten sailed to America because he was hungry for gold after having been kept on a short financial leash by his father sounds more plausible than to merely say that Hutten sailed to America because he was hungry for gold. When we are dealing with a story, the greater number of causes does not detract from our sense of understanding, but it enhances it.

This sense of greater understanding is, however, irrational. The story that Hutten sailed to America because he was hungry for gold is actually more likely to be true than the story that Hutten sailed to America because he was hungry for gold after having been kept on a short financial leash by his father. By simple statistical logic, the compound probability of two events cannot be greater than the probability of either component (Dawes, 1988).

Looking back, our causal, or reverse, inferences tend to be more fallible than our forward inferences, which we make when we look to the future. When we fail to understand this asymmetry, we are committing a reverse inference fallacy. The historical analysis of text, whenever it becomes interpretational, is vulnerable to this fallacy. Yet, I think it would be a mistake to strip all interpretation from our efforts to make sense of the past, and settle for a Spartan variety of pure description. Naturally, psychological questions will be asked when we contemplate the motives of actors described in the text. My suggestion is that we proceed with caution, and that we let theory and research in experimental psychology be our guide. When we do this, we will also learn where our own judgment is prone to bias and error, and this awareness will allow us to keep an open mind for alternative readings of the text.

Experimental psychology can help us steer clear of the text fallacy. By text fallacy I mean the idea that everything in a text (written or otherwise) is meaningful, and that our task is to achieve a full interpretation that takes each and every detail into account. The statistical analog of the text fallacy is the overfitting of empirical data. Suppose you are looking at a graph of the daily high temperatures in Santiago. You can certainly fit a mathematical function that connects all the 365 data points, but you may need 365 parameters to do it. This is overfitting (Hoffrage & Hertwig, 2012). The mathematical solution is perfect, but only with respect to the past. It is of little use when predicting the temperatures for the next year. In contrast, you can fit the data with a simple sinusoid function that captures the gentle rise and fall of temperature over the seasons. This function uses only few parameters, while treating day-to-day variation as statistical error. Because of this, a simple function does a good job predicting next year’s temperatures. And by the way, it is also beautiful.

A deeper layer of the text fallacy is the idea that the true meaning of text is never apparent, but encrypted. From the Talmudic tradition of interpreting the Torah to the Freudian tradition of interpreting dreams, we see how an analytical mindset can mystify rather than clarify. The core of this cryptomanic fallacy is the premise that no detail, however small, is a matter of randomness or error. The idea of randomness is anathema to the goal of coherent story construction. Randomness seems like waste. It violates the idea that stories are created by intelligent design. The good story presumes an intelligent story-teller. The intelligent story-teller makes no errors. He has edited his text to purge all redundancy, contradiction, and waste. Given the premise that a story, or text, is a well-crafted construction, we by necessity become increasingly confident as our story of the story becomes more subtle, more complex, and, alas, more improbable.

In the psychology of judgment and decision-making, the error of overconfidence is the cardinal mistake of the thinking mind. It is the Ur-error. Oskamp (1965) showed that clinical psychologists become increasingly confident about their case interpretations as they gather more information, while the accuracy of their judgments actually goes down (see Moore & Healy, 2008, for a contemporary view).

Let’s review: Human history is populated by individuals who did stuff. They come to us wrapped up and dressed in text. We want to know – among other things – why they did what they did, in other words, we are asking causal questions. I have claimed, and shown by example, that experimental psychology has a contribution to make in this enterprise. Despite the hazards of reverse inference, we can put forth conjectures about the human element of our history without relying on the popular but sterile concepts of psychoanalysis. I am proposing a win-win scenario. Experimental psychology can gain by showing that some of her lessons are applicable to the humanities, and the humanities can, by making contact with experimental science, see their interpretations become more general and credible . . . if that’s what they want. I see this proposal as being in the spirit of Alexander von Humboldt, who as a scientist was a unifier. He viewed nature as an organic whole and he used all manner of method at his disposal.

In a keynote address to the VII Humboldt conference, Professor Ottmar Ette talked about Humboldt’s epistemic revolution. He described Humboldt’s style as “nomadic instead of monadic,” and he noted that “Art brings together what science divides.” In this spirit, I think that experimental psychology can make a contribution to the scholarship of text. Alternatively, as one symposiast put it to me in conversation, you may assert that “Science is a myth.” She appeared to have missed the irony of her making this statement during a conference on Humboldt.

Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. Organizational Behavior and Human Decision Processes, 35, 124-140.

Bicchieri, C. (2006). The grammar of society: The nature and dynamics of social norms. New York, NY: Cambridge University Press.

Cox, D. R., & Wermuth, N. (2001). Causal inference and statistical fallacies. International Encyclopedia of the Social and Behavioral Sciences, 1554-1561.

Dawes, R. M. (1988). Rational choice in an uncertain world. San Diego, CA: Harcourt Brace Jovanovich.

Fehr, E., & Fischbacher, U. (2004). Third-party punishment and social norms. Evolution and Human Behavior, 25, 63-87.

Gilmore, D. D. (1990). Manhood in the making: Cultural concepts of masculinity. New Haven, CT: Yale University Press.

Hoffrage, U., & Hertwig, R. (2012). Simple heuristics in a complex social world. In J. I. Krueger (ed.), Social judgment and decision making (pp. 135-150). New York, NY: Psychology Press.

Krueger, J. I. (2013). Psychology of a conquistador. Jahrbuch für Europäische Überseegeschichte, 13, 137-145. Leipzig, Germany.

Krueger, J. I. (2013). Psychology of a conquistador. Jahrbuch für Europäische Überseegeschichte [Annals of European Overseas History], 13, 37-41. Harrassowitz Verlag, Wiesbaden, Germany.

Leung, A. K.-Y., & Cohen, D. (2011). Within- and between-culture variation: Individual differences and the cultural logics of honor, face, and dignity cultures. Journal of Personality and Social Psychology, 100, 507-526.

Moore, D., & Healy, P. J. (2008). The trouble with overconfidence. Psychological Review, 115, 502-517.

Oskamp, S. (1965). Overconfidence in case-study judgments. Journal of Consulting Psychology, 29, 261-265.

Schmitt, E., & Simmer, G. (1999). Tod am Tocuyo: Die Suche nach den Hintergründen der Ermordung Philipps von Hutten 1541 – 1550 (Death on the Tocuyo: The search for the reasons behind the murder of Philipp von Hutten 1541 – 1550). Berlin: Spitz.

Schmitt, E., & von Hutten, F. K. (1996). Das Gold der Neuen Welt: Die Papiere des Welser-Konquistadors und Generalkapitäns von Venezuela Philipp von Hutten 1534 – 1541 (The gold of the New World: The papers of the Welser conquistator and captain general of Venezuela Philipp von Hutten). Hildburghausen: Verlap Frankenschwelle.

Note. This post is based on a talk entitled Conquest, Confidence, and Text: Philipp von Hutten and the case for psychology, prepared for the VIIth International and Interdisciplinary Commemorative Congress in Honor of Alexander von Humboldt, Claudio Gay, and Ignacio Domeyko, Santiago, Chile, January 5 – 10, 2014.