Rational life plans and the stopping problem

Image: a poor solution to the stopping problem

In earlier posts I discussed the question of "rational plans of life" (link, link, link, link) and argued that standard theories of rational decision making under uncertainty don't do well in this context. I argued instead that rationality in navigating and building a life is not analogous to remodeling your kitchen; instead, it involves provisional clarification of the goals and values that one embraces, and then a kind of step-by-step, self-critical direction-setting in the choices that one makes over time in ways that honor these values.

Brian Christian and Tom Griffiths' Algorithms to Live By: The Computer Science of Human Decisions provides a very interesting additional perspective on this problem of living a life. The authors describe the algorithms that computer science has discovered to handle difficult choice problems, and they make an effort to both explain (generally) how the problem is solved formally and how it finds application in ordinary situations of human decision-making over an extended time -- such as the challenging question of where to stop for a meal on a long road trip, or which candidate to hire as an executive assistant.

The key features of decision-making that drive much of their discussion are time and uncertainty. We often have to make decisions and choices among options where we do not know the qualities of the items on offer (restaurants to consider for a special meal, individuals who are prospective friends, who to hire for an important position), the likelihood of success of a given item, and where we often cannot return to a choice we've already rejected. (If we are driving between Youngstown and Buffalo there are only finitely many restaurants where we might stop for a meal; but once we've passed New Bangkok Restaurant at exit 50 on the interstate, we are unlikely to return when we haven't found a better choice by exit 55.)

The stopping problem seems relevant to the problem of formulating a rational plan of life, since the stream of life events and choices in a person's life is one-directional, and it is rare to be able to return to an option that was rejected at a prior moment. In hindsight -- should I have gone to Harvard for graduate school, or would Cornell or Princeton have been a better choice? The question is literally pointless; it cannot be undone. Life, like history, proceeds in only one direction. Many life choices must be made before a full comparison of the quality of the options and the consequences of one choice or another can be fully known. And waiting until all options have been reviewed often means that the earlier options are no longer available -- just like that Thai restaurant on the Ohio Turnpike at exit 50.

The algorithms that surround the stopping problem have a specific role in decision-making in ordinary life circumstances: we will make better decisions under conditions of uncertainty and irreversibility if we understand something about the probabilities of the idea that "a better option is still coming up". We need to have some intuitive grasp of the dialectic of "exploration / exploitation" that the stopping problem endorses. As Christian and Griffiths put it, "exploration is gathering information, and exploitation is using the information you have to get a known good result" (32). How long should we continue to gather information (exploration) and at what point should we turn to active choice ("choose the next superior candidate that comes along")? If a person navigates life by exploring 90% of options before choosing, he or she is likely to do worse than less conservative decision-makers; but likewise about the person who chooses after seeing 5% of the options.

There is a very noticeable convergence between the algorithms of stopping and Herbert Simon's theory of satisficing (link). (The authors note this parallel in a footnote.) Simon noted that the heroic assumptions of economic rationality are rarely satisfied in actual human decision-making: full information about the probabilities and utilities associated with a finite range of outcomes, and choice guided by choosing that option with the greatest expected utility. He notes that this view of rationality requires an unlimited budget for information gathering, and that -- at some point -- the cost of further search outweighs the probably gains of finding the optimal solution. Simon too argues that rational decision-makers "stop" in their choices: they set a threshold value for quality and value, initiate a search, and select the first option that meets the threshold. "Good enough" beats "best possible". If I decide I need a pair of walking shoes, I decide on price and quality -- less than $100, all leather, good tread, comfortable fit -- and I visit a sequence of shoe stores, with the plan of buying the first pair of shoes that meets the threshold. But the advantage of the search algorithm described here is that it does not require a fixed threshold in advance, and it would appear to give a higher probability of making the best possible choice among all available options. As a speculative guess, it seems as though searches guided by a fixed threshold would score lower over time than searches guided by a balanced "explore, then exploit" strategy, without the latter being overwhelmed by information costs.

In one of the earlier posts on "rational life plans" I suggested that rationality comes into life-planning in several different ways:

We might describe this process as one that involves local action-rationality guided by medium term strategies and oriented towards long term objectives. Rationality comes into the story at several points: assessing cause and effect, weighing the importance of various long term goals, deliberating across conflicting goals and values, working out the consequences of one scenario or another, etc. (link)

The algorithms of stopping are clearly relevant to the first part of the story -- local action-rationality. It is not so clear that the stopping problem arises in the same way in the other two levels of life-planning rationality. Deliberation about longterm objectives is not sequential in the way that deciding about which highway exit to choose for supper is; rather, the deliberating individual can canvas a number of objectives simultaneously and make deliberative choices among them. And choosing medium-term strategies seems to have a similar temporal logic: identify a reasonable range of possible strategies, compare their strengths and weaknesses, and choose the best. So the stopping problem seems to be relevant to the implementation phase of living, not the planning and projecting parts. We don't need the stopping algorithm to decide to visit the grandchildren in Scranton, or in deciding which route across the country to choose for the long drive; but we do need it for deciding the moment-to-moment options that arise -- which hotel, which restaurant, which stretch of beach, which tourist attraction to visit along the way. This seems to amount to a conclusion: the stopping problem is relevant to a certain class of choices that come as an irreversible series, but not relevant to deliberation among principles, values, or guiding goals.

(Christian and Griffiths describe the results of research on the stopping problem; but the book does not give a clear description of how the math works. Here is a somewhat more detailed explanation of the solution to the stopping problem in American Scientist; link. Essentially the solution -- wait and observe for the first 37% of options, then taken the next option better than any of those seen to date -- follows from a calculation of the probability of the distribution of "best choices" across the random series of candidates. And it can be proven that both lower and higher thresholds -- less exploration or more exploration -- lead to lower average payoffs.)

Slime mold intelligence

We often think of intelligent action in terms of a number of ideas: goal-directedness, belief acquisition, planning, prioritization of needs and wants, oversight and management of bodily behavior, and weighting of risks and benefits of alternative courses of action. These assumptions presuppose the existence of the rational subject who actively orchestrates goals, beliefs, and priorities into an intelligent plan of action. (Here is a series of posts on "rational life plans"; link, link, link.)

It is interesting to discover that some simple adaptive systems apparently embody an ability to modify behavior so as to achieve a specific goal without possessing a number of these cognitive and computational functions. These systems seem to embody some kind of cross-temporal intelligence. An example that is worth considering is the spatial and logistical capabilities of the slime mold. A slime mold is a multi-cellular "organism" consisting of large numbers of independent cells without a central control function or nervous system. It is perhaps more accurate to refer to the population as a colony rather than an organism. Nonetheless the slime mold has a remarkable ability to seek out and "optimize" access to food sources in the environment through the creation of a dynamic network of tubules established through space.

The slime mold lacks beliefs, it lacks a central cognitive function or executive function, it lacks "memory" -- and yet the organism (colony?) achieves a surprising level of efficiency in exploring and exploiting the food environment that surrounds it. Researchers have used slime molds to simulate the structure of logistical networks (rail and road networks, telephone and data networks), and the results are striking. A slime mold colony appear to be "intelligent" in performing the task of efficiently discovering and exploiting food sources in the environment in which it finds itself.

One of the earliest explorations of this parallel between biological networks and human-designed networks was Tero et al, "Rules for Biologically Inspired Adaptive Network Design" in Science in 2010 (link). Here is the abstract of their article:

Abstract Transport networks are ubiquitous in both social and biological systems. Robust network performance involves a complex trade-off involving cost, transport efficiency, and fault tolerance. Biological networks have been honed by many cycles of evolutionary selection pressure and are likely to yield reasonable solutions to such combinatorial optimization problems. Furthermore, they develop without centralized control and may represent a readily scalable solution for growing networks in general. We show that the slime mold Physarum polycephalum forms networks with comparable efficiency, fault tolerance, and cost to those of real-world infrastructure networks—in this case, the Tokyo rail system. The core mechanisms needed for adaptive network formation can be captured in a biologically inspired mathematical model that may be useful to guide network construction in other domains.

Their conclusion is this:

Overall, we conclude that the Physarum networks showed characteristics similar to those of the [Japanese] rail network in terms of cost, transport efficiency, and fault tolerance. However, the Physarum networks self-organized without centralized control or explicit global information by a process of selective reinforcement of preferred routes and simultaneous removal of redundant connections. (441)

They attempt to uncover the mechanism through which this selective reinforcement of routes takes place, using a simulation "based on feedback loops between the thickness of each tube and internal protoplasmic flow in which high rates of streaming stimulate an increase in tube diameter, whereas tubes tend to decline at low flow rates" (441). The simulation is successful in approximately reproducing the observable dynamics of evolution of the slime mold networks. Here is their summary of the simulation:

Our biologically inspired mathematical model can capture the basic dynamics of network adaptability through iteration of local rules and produces solutions with properties comparable or better than those real-world infrastructure networks. Furthermore, the model has a number of tunable parameters that allow adjustment of the benefit-cost ratio to increase specific features, such as fault tolerance or transport efficiency, while keeping costs low. Such a model may provide a useful starting point to improve routing protocols and topology control for self-organized networks such as remote sensor arrays, mobile ad hoc networks, or wireless mesh networks. (442)

Here is a summary description of what we might describe as the "spatial problem-solving abilities" of the slime mold based on this research by Katherine Harman in a Scientific American blog post (link):

Like the humans behind a constructed network, the organism is interested in saving costs while maximizing utility. In fact, the researchers wrote that this slimy single-celled amoeboid can "find the shortest path through a maze or connect different arrays of food sources in an efficient manner with low total length yet short average minimum distances between pairs of food sources, with a high degree of fault tolerance to accidental disconnection"—and all without the benefit of "centralized control or explicit global information." In other words, it can build highly efficient connective networks without the help of a planning board.

This research has several noteworthy features. First, it seems to provide a satisfactory account of the mechanism through which slime mold "network design intelligence" is achieved. Second, the explanation depends only on locally embodied responses at the local level, without needing to appeal to any sort of central coordination or calculation. The process is entirely myopic and locally embodied, and the "global intelligence" of the colony is entirely generated by the locally embodied action states of the individual mold cells. And finally, the simulation appears to offer resources for solving real problems of network design, without the trouble of sending out a swarm of slime mold colonies to work out the most efficient array of connectors.

We might summarize this level of slime-mold intelligence as being captured by:

• trial-and-error extension of lines of exploration
• localized feedback on results of a given line leading to increase/decrease of the volume of that line

This system is decentralized and myopic with no ability to plan over time and no "over-the-horizon" vision of potential gains from new lines of exploration. In these respects slime-mold intelligence has a lot in common with the evolution of species in a given ecological environment. It is an example of "climbing Mt. Improbable" involving random variation and selection based on a single parameter (volume of flow rather than reproductive fitness). If this is a valid analogy, then we might be led to expect that the slime mold is capable of finding local optima in network design but not global optima. (Or the slime colony may avoid this trap by being able to fully explore the space of network configurations over time.) What the myopia of this process precludes is the possibility of strategic action and planning -- absorbing sacrifices at an early part of the process in order to achieve greater gains later in the process. Slime molds would not be very good at chess, Go, or war.

I've been tempted to offer the example of slime mold intelligence as a description of several important social processes apparently involving collective intentionality: corporate behavior and discovery of pharmaceuticals (link) and the aggregate behavior of large government agencies (link).

On pharmaceutical companies:

So here's the question for consideration here: what if we attempted to model the system of population, disease, and the pharmaceutical industry by representing pharma and its multiple research and discovery units as the slime organism and the disease space as a set of disease populations with different profitability characteristics? Would we see a major concentration of pharma slime around a few high-frequency, high profit disease-drug pairs? Would we see substantial under-investment of pharma slime on low frequency low profit "orphan" disease populations? And would we see hyper-concentrations around diseases whose incidence is responsive to marketing and diagnostic standards? (link)

On the "intelligence" of firms and agencies:

But it is perfectly plain that the behavior of functional units within agencies are only loosely controlled by the will of the executive. This does not mean that executives have no control over the activities and priorities of subordinate units. But it does reflect a simple and unavoidable fact about large organizations. An organization is more like a slime mold than it is like a control algorithm in a factory. (link)

In each instance the analogy works best when we emphasize the relative weakness of central strategic control (executives) and the solution-seeking activities of local units. But of course there is a substantial degree of executive involvement in both private and public organizations -- not fully effective, not algorithmic, but present nonetheless. So the analogy is imperfect. It might be more accurate to say that the behavior of large complex organizations incorporates both imperfect central executive control and the activities of local units with myopic search capabilities coupled with feedback mechanisms. The resulting behavior of such a system will not look at all like the idealized business-school model of "fully implemented rational business plans", but it will also not look like a purely localized resource-maximizing network of activities.

******

Here is a very interesting set of course notes in which Prof. Donglei Du from the University of New Brunswick sets the terms for a computational and heuristic solution to a similar set of logistics problems. Du asks his students to consider the optimal locations of warehouses to supply retailers in multiple locations; link. Here is how Du formulates the problem:

*     Assuming that plants and retailer locations are fixed, we concentrate on the following strategic decisions in terms of warehouses.

• Pick the optimal number, location, and size of warehouses 
• Determine optimal sourcing strategy
• Which plant/vendor should produce which product 
• Determine best distribution channels
• Which warehouses should service which retailers

The objective is to design or reconfigure the logistics network so as to minimize annual system-wide costs, including

• Production/ purchasing costs
• Inventory carrying costs, and facility costs (handling and fixed costs)
• Transportation costs

As Du demonstrates, the mathematics involved in an exact solution are challenging, and become rapidly more difficult as the number of nodes increases.

Even though this example looks rather similar to the rail system example above, it is difficult to see how it might be modeled using a slime mold colony. The challenge seems to be that the optimization problem here is the question of placement of nodes (warehouses) rather than placement of routes (tubules).

Herbert Simon's theories of organizations

Image: detail from Family Portrait 2 1965 

(Creative Commons license, Richard Rappaport)

Herbert Simon made paradigm-changing contributions to the theory of rational behavior, including particularly his treatment of "satisficing" as an alternative to "maximizing" economic rationality (link). It is therefore worthwhile examining his views of organizations and organizational decision-making and action -- especially given how relevant those theories are to my current research interest in organizational dysfunction. His highly successful book Administrative Behavior went through four editions between 1947 and 1997 -- more than fifty years of thinking about organizations and organizational behavior. The more recent editions consist of the original text and "commentary" chapters that Simon wrote to incorporate more recent thinking about the content of each of the chapters.

Here I will pull out some of the highlights of Simon's approach to organizations. There are many features of his analysis of organizational behavior that are worth noting. But my summary assessment is that the book is surprisingly positive about the rationality of organizations and the processes through which they collect information and reach decisions. In the contemporary environment where we have all too many examples of organizational failure in decision-making -- from Boeing to Purdue Pharma to the Federal Emergency Management Agency -- this confidence seems to be fundamentally misplaced. The theorist who invented the idea of imperfect rationality and satisficing at the individual level perhaps should have offered a somewhat more critical analysis of organizational thinking.

The first thing that the reader will observe is that Simon thinks about organizations as systems of decision-making and execution. His working definition of organization highlights this view:

In this book, the term organization refers to the pattern of communications and relations among a group of human beings, including the processes for making and implementing decisions. This pattern provides to organization members much of the information and many of the assumptions, goals, and attitudes that enter into their decisions, and provides also a set of stable and comprehensible expectations as to what the other members of the group are doing and how they will react to what one says and does. (18-19).

What is a scientifically relevant description of an organization? It is a description that, so far as possible, designates for each person in the organization what decisions that person makes, and the influences to which he is subject in making each of these decisions. (43)

The central theme around which the analysis has been developed is that organization behavior is a complex network of decisional processes, all pointed toward their influence upon the behaviors of the operatives -- those who do the action 'physical' work of the organization. (305)

The task of decision-making breaks down into the assimilation of relevant facts and values -- a distinction that Simon attributes to logical positivism in the original text but makes more general in the commentary. Answering the question, "what should we do?", requires a clear answer to two kinds of questions: what values are we attempting to achieve? And how does the world work such that interventions will bring about those values?

It is refreshing to see Simon's skepticism about the "rules of administration" that various generations of organizational theorists have advanced -- "specialization," "unity of command," "span of control," and so forth. Simon describes these as proverbs rather than as useful empirical discoveries about effective administration. And he finds the idea of "schools of management theory" to be entirely unhelpful (26). Likewise, he is entirely skeptical about the value of the economic theory of the firm, which abstracts from all of the arrangements among participants that are crucial to the internal processes of the organization in Simon's view. He recommends an approach to the study of organizations (and the design of organizations) that focuses on the specific arrangements needed to bring factual and value claims into a process of deliberation leading to decision -- incorporating the kinds of specialization and control that make sense for a particular set of business and organizational tasks.

An organization has only two fundamental tasks: decision-making and "making things happen". The decision-making process involves intelligently gathering facts and values and designing a plan. Simon generally approaches this process as a reasonably rational one. He identifies three kinds of limits on rational decision-making:

• The individual is limited by those skills, habits, and reflexes which are no longer in the realm of the conscious...
• The individual is limited by his values and those conceptions of purpose which influence him in making his decision...
• The individual is limited by the extent of his knowledge of things relevant to his job. (46)

And he explicitly regards these points as being part of a theory of administrative rationality:

Perhaps this triangle of limits does not completely bound the area of rationality, and other sides need to be added to the figure. In any case, the enumeration will serve to indicate the kinds of considerations that must go into the construction of valid and noncontradictory principles of administration. (47)

The "making it happen" part is more complicated. This has to do with the problem the executive faces of bringing about the efficient, effective, and loyal performance of assigned tasks by operatives. Simon's theory essentially comes down to training, loyalty, and authority.

If this is a correct description of the administrative process, then the construction of an efficient administrative organization is a problem in social psychology. It is a task of setting up an operative staff and superimposing on that staff a supervisory staff capable of influencing the operative group toward a pattern of coordinated and effective behavior. (2)

To understand how the behavior of the individual becomes a part of the system of behavior of the organization, it is necessary to study the relation between the personal motivation of the individual and the objectives toward which the activity of the organization is oriented. (13-14) 

Simon refers to three kinds of influence that executives and supervisors can have over "operatives": formal authority (enforced by the power to hire and fire), organizational loyalty (cultivated through specific means within the organization), and training. Simon holds that a crucial role of administrative leadership is the task of motivating the employees of the organization to carry out the plan efficiently and effectively.

Later he refers to five "mechanisms of organization influence" (112): specialization and division of task; the creation of standard practices; transmission of decisions downwards through authority and influence; channels of communication in all directions; and training and indoctrination. Through these mechanisms the executive seeks to ensure a high level of conformance and efficient performance of tasks.

What about the actors within an organization? How do they behave as individual actors? Simon treats them as "boundedly rational":

To anyone who has observed organizations, it seems obvious enough that human behavior in them is, if not wholly rational, at least in good part intendedly so. Much behavior in organizations is, or seems to be, task-oriented--and often efficacious in attaining its goals. (88)

But this description leaves out altogether the possibility and likelihood of mixed motives, conflicts of interest, and intra-organizational disagreement. When Simon considers the fact of multiple agents within an organization, he acknowledges that this poses a challenge for rationalistic organizational theory:

Complications are introduced into the picture if more than one individual is involved, for in this case the decisions of the other individuals will be included among the conditions which each individual must consider in reaching his decisions. (80)

This acknowledges the essential feature of organizations -- the multiplicity of actors -- but fails to treat it with the seriousness it demands. He attempts to resolve the issue by invoking cooperation and the language of strategic rationality: "administrative organizations are systems of cooperative behavior. The members of the organization are expected to orient their behavior with respect to certain goals that are taken as 'organization objectives'" (81). But this simply presupposes the result we might want to occur, without providing a basis for expecting it to take place.

With the hindsight of half a century, I am inclined to think that Simon attributes too much rationality and hierarchical purpose to organizations.

The rational administrator is concerned with the selection of these effective means. For the construction of an administrative theory it is necessary to examine further the notion of rationality and, in particular, to achieve perfect clarity as to what is meant by "the selection of effective means." (72)  

These sentences, and many others like them, present the task as one of defining the conditions of rationality of an organization or firm; this takes for granted the notion that the relations of communication, planning, and authority can result in a coherent implementation of a plan of action. His model of an organization involves high-level executives who pull together factual information (making use of specialized experts in this task) and integrating the purposes and goals of the organization (profits, maintaining the health and safety of the public, reducing poverty) into an actionable set of plans to be implemented by subordinates. He refers to a "hierarchy of decisions," in which higher-level goals are broken down into intermediate-level goals and tasks, with a coherent relationship between intermediate and higher-level goals. "Behavior is purposive in so far as it is guided by general goals or objectives; it is rational in so far as it selects alternatives which are conducive to the achievement of the previously selected goals" (4).  And the suggestion is that a well-designed organization succeeds in establishing this kind of coherence of decision and action.

It is true that he also asserts that decisions are "composite" --

It should be perfectly apparent that almost no decision made in an organization is the task of a single individual. Even though the final responsibility for taking a particular action rests with some definite person, we shall always find, in studying the manner in which this decision was reached, that its various components can be traced through the formal and informal channels of communication to many individuals ... (305)

But even here he fails to consider the possibility that this compositional process may involve systematic dysfunctions that require study. Rather, he seems to presuppose that this composite process itself proceeds logically and coherently. In commenting on a case study by Oswyn Murray (1923) on the design of a post-WWI battleship, he writes: "The point which is so clearly illustrated here is that the planning procedure permits expertise of every kind to be drawn into the decision without any difficulties being imposed by the lines of authority in the organization" (314). This conclusion is strikingly at odds with most accounts of science-military relations during World War II in Britain -- for example, the pernicious interference of Frederick Alexander Lindemann with Patrick Blackett over Blackett's struggles to create an operations-research basis for anti-submarine warfare (Blackett's War: The Men Who Defeated the Nazi U-Boats and Brought Science to the Art of Warfare). His comments about the processes of review that can be implemented within organizations (314 ff.) are similarly excessively optimistic -- contrary to the literature on principal-agent problems in many areas of complex collaboration.

This is surprising, given Simon's contributions to the theory of imperfect rationality in the case of individual decision-making. Against this confidence, the sources of organizational dysfunction that are now apparent in several literatures on organization make it more difficult to imagine that organizations can have a high success rate in rational decision-making. If we were seeking for a Simon-like phrase for organizational thinking to parallel the idea of satisficing, we might come up with the notion of "bounded localistic organizational rationality": "locally rational, frequently influenced by extraneous forces, incomplete information, incomplete communication across divisions, rarely coherent over the whole organization".

Simon makes the point emphatically in the opening chapters of the book that administrative science is an incremental and evolving field. And in fact, it seems apparent that his own thinking continued to evolve. There are occasional threads of argument in Simon's work that seem to point towards a more contingent view of organizational behavior and rationality, along the lines of Fligstein and McAdam's theories of strategic action fields. For example, when discussing organizational loyalty Simon raises the kind of issue that is central to the strategic action field model of organizations: the conflicts of interest that can arise across units (11). And in the commentary on Chapter I he points forward to the theories of strategic action fields and complex adaptive systems:

The concepts of systems, multiple constituencies, power and politics, and organization culture all flow quite naturally from the concept of organizations as complex interactive structures held together by a balance of the inducements provided to various groups of participants and the contributions received from them. (27)

The book has been a foundational contribution to organizational studies. At the same time, if Herbert Simon were at the beginning of his career and were beginning his study of organizational decision-making today, I suspect he might have taken a different tack. He was plainly committed to empirical study of existing organizations and the mechanisms through which they worked. And he was receptive to the ideas surrounding the notion of imperfect rationality. The current literature on the sources of contention and dysfunction within organizations (Perrow, Fligstein, McAdam, Crozier, ...) might well have led him to write a different book altogether, one that gave more attention to the sources of failures of rational decision-making and implementation alongside the occasional examples of organizations that seem to work at a very high level of rationality and effectiveness.

Declining industries

Why is it so difficult for leaders in various industries and sectors to seriously address the existential threats that sometimes arise? Planning for marginal changes in the business environment is fairly simple; problems can be solved, costs can be cut, and the firm can stay in the black. But how about more radical but distant threats? What about the grocery sector when confronted by Amazon's radical steps in food selling? What about Polaroid or Kodak when confronted by the rise of digital photography in the 1990s? What about the US steel industry in the 1960s when confronted with rising Asian competition and declining manufacturing facilities?

From the outside these companies and sectors seem like dodos incapable of confronting the threats that imperil them. They seem to be ignoring oncoming train wrecks simply because these catastrophes are still in the distant future. And yet the leaders in these companies were generally speaking talented, motivated men and women. So what are the organizational or cognitive barriers that arise to make it difficult for leaders to successfully confront the biggest threats they face?

Part of the answer seems to be the fact that distant hazards seem smaller than the more immediate and near-term challenges that an organization must face; so there is a systematic bias towards myopic decision-making. This sounds like a Kahneman-Tversky kind of cognitive shortcoming.

A second possible explanation is that it is easy enough to persuade oneself that distant threats will either resolve themselves organically or that the organization will discover novel solutions in the future. This seems to be part of the reason that climate-change foot-draggers take the position they do: that "things will sort out", "new technologies will help solve the problems in the future." This sounds like a classic example of weakness of the will -- an unwillingness to rationally confront hard truths about the future that ought to influence choices today but often don't.

Then there is the timeframe of accountability that is in place in government, business, and non-profit organizations alike. Leaders are rewarded and punished for short-term successes and failures, not prudent longterm planning and preparation. This is clearly true for term-limited elected officials, but it is equally true for executives whose stakeholders evaluate performance based on quarterly profits rather than longterm objectives and threats.

We judge harshly those leaders who allow their firms or organizations to perish because of a chronic failure to plan for substantial change in the environments in which they will need to operate in the future. Nero is not remembered kindly for his dedication to his fiddle. And yet at any given time, many industries are in precisely that situation. What kind of discipline and commitment can protect organizations against this risk?

This is an interesting question in the abstract. But it is also a challenging question for people who care about the longterm viability of colleges and universities. Are there forces at work today that will bring about existential crisis for universities in twenty years (enrollments, tuition pressure, technology change)? Are there technological or organizational choices that should be made today that would help to avert those crises in the future? And are university leaders taking the right steps to prepare their institutions for the futures they will face in several decades?

More on cephalopod minds

When I first posted on cephalopod intelligence a year or so ago, I assumed it would be a one-off diversion into the deep blue sea (link). But now I've read the fascinating recent book by Peter Godfrey-Smith, Other Minds: The Octopus, the Sea, and the Deep Origins of Consciousness, and it is interesting enough to justify a second deep dive. Godfrey-Smith is a philosopher, but he is also a scuba diver, and his interest in cephalopods derives from his experiences under water. This original stimulus has led to two very different lines of inquiry. What is the nature of the mental capacities of an octopus? And how did "intelligence" happen to evolve twice on earth through such different pathways? Why is a complex nervous system an evolutionary advantage for a descendent of a clam?

Both questions are of philosophical interest. The nature of consciousness, intelligence, and reasoning has been of great concern to philosophers in the study of the philosophy of mind. The questions that arise bring forth a mixture of difficult conceptual, empirical, and theoretical issues: how does consciousness relate to behavioral capacity? Are intelligence and consciousness interchangeable? What evidence would permit us to conclude that a given species of animal has consciousness and reasoning ability?

The evolutionary question is also of interest to philosophers. The discipline of the philosophy of biology focuses much of its attention on the issues raised by evolutionary theory. Elliott Sober's work illustrates this form of philosophical thinking -- for example, The Nature of Selection: Evolutionary Theory in Philosophical Focus, Evidence and Evolution: The Logic Behind the Science. Godfrey-Smith tells an expert's story of the long evolution of mollusks, in and out of their shells, with emerging functions and organs well suited to the opportunities available in their oceanic environments. One of the evolutionary puzzles to be considered is the short lifespan of octopuses and squid -- just a few years (160). Why would the organism invest so heavily in a cognitive system that supported its life for such a short time?

A major part of the explanation that G-S favors involves the fact that octopuses are hunters, and a complex nervous system is more of an advantage for predator than prey. (Wolves are more intelligent than elk, after all!) Having a nervous system that supports anticipation, planning, and problem solving turns out to be an excellent preparation for being a predator. Here is a good example of how that cognitive advantage plays out for the octopus:

David Scheel, who works mostly with the giant Pacific octopus, feeds his animals whole clams, but as his local animals in Prince William Sound do not routinely eat clams, he has to teach them about the new food source. So he partly smashes a clam and gives it to the octopus. Later, when he gives the octopus an intact clam, the octopus knows that it’s food, but does not know how to get at the meat. The octopus will try all sorts of methods, drilling the shell and chipping the edges with its beak, manipulating it in every way possible … and then eventually it learns that its sheer strength is sufficient: if it tries hard enough, it can simply pull the shell apart. (70)

Exploration, curiosity, experimentation, and play are crucial components of the kind of flexibility that organisms with big nervous systems bring to earning their living.

G-S brings up a genuinely novel aspect of the organismic value of a complex nervous system: not just problem-solving applied to the external environment, but coordination of the body itself. Intelligence evolves to handle the problem of coordinating the motions of the parts of the body.

The cephalopod body, and especially the octopus body, is a unique object with respect to these demands. When part of the molluscan “foot” differentiated into a mass of tentacles, with no joints or shell, the result was a very unwieldy organ to control. The result was also an enormously useful thing, if it could be controlled. The octopus’s loss of almost all hard parts compounded both the challenge and the opportunities. A vast range of movements became possible, but they had to be organized, had to be made coherent. Octopuses have not dealt with this challenge by imposing centralized governance on the body; rather, they have fashioned a mixture of local and central control. One might say the octopus has turned each arm into an intermediate-scale actor. But it also imposes order, top-down, on the huge and complex system that is the octopus body. (71)

In this picture, neurons first multiply because of the demands of the body, and then sometime later, an octopus wakes up with a brain that can do more. (72)

This is a genuinely novel and intriguing idea about the creation of a new organism over geological time. It is as if a plastic self-replicating and self-modifying artifact bootstrapped itself from primitive capabilities into a directed and cunning predator. Or perhaps it is a preview of the transition that artificial intelligence systems embodying adaptable learning processes and expanding linkages to the control systems of the physical world may take in the next fifty years.  

What about the evolutionary part of the story? Here is a short passage where Godfrey-Smith considers the long evolutionary period that created both vertebrates and mollusks:

The history of large brains has, very roughly, the shape of a letter Y. At the branching center of the Y is the last common ancestor of vertebrates and mollusks. From here, many paths run forward, but I single out two of them, one leading to us and one to cephalopods. What features were present at that early stage, available to be carried forward down both paths? The ancestor at the center of the Y certainly had neurons. It was probably a worm-like creature with a simple nervous system, though. It may have had simple eyes. Its neurons may have been partly bunched together at its front, but there wouldn’t have been much of a brain there. From that stage the evolution of nervous systems proceeds independently in many lines, including two that led to large brains of different design. (65)

The primary difference that G-S highlights here is the nature of the neural architecture that each line eventually favors: a central cord connecting periphery to a central brain; and a decentralized network of neurons distributed over the whole body.

Further, much of a cephalopod’s nervous system is not found within the brain at all, but spread throughout the body. In an octopus, the majority of neurons are in the arms themselves— nearly twice as many as in the central brain. The arms have their own sensors and controllers. They have not only the sense of touch, but also the capacity to sense chemicals— to smell, or taste. Each sucker on an octopus’s arm may have 10,000 neurons to handle taste and touch. Even an arm that has been surgically removed can perform various basic motions, like reaching and grasping. (67)

So what about the "alien intelligence" part of G-S's story? G-S emphasizes the fact that octopus mentality is about as alien to human experience and evolution as it could be.

Cephalopods are an island of mental complexity in the sea of invertebrate animals. Because our most recent common ancestor was so simple and lies so far back, cephalopods are an independent experiment in the evolution of large brains and complex behavior. If we can make contact with cephalopods as sentient beings, it is not because of a shared history, not because of kinship, but because evolution built minds twice over. This is probably the closest we will come to meeting an intelligent alien. (9)

This too is intriguing. G-S is right: the evolutionary story he works through here gives great encouragement for the idea that an organism in a complex environment and a few bits of neuronal material can evolve in wildly different pathways, leading to cognitive capabilities and features of awareness that are dramatically different from human intelligence. Life is plastic and evolutionary time is long. The ideas of the unity of consciousness and the unified self don't have any particular primacy or uniqueness. For example: 

The octopus may be in a sort of hybrid situation. For an octopus, its arms are partly self—they can be directed and used to manipulate things. But from the central brain’s perspective, they are partly non-self too, partly agents of their own. (103)

So there is nothing inherently unique about human intelligence, and no good reason to assume that all intelligent creatures would find a basis for mutual understanding and communication. Sorry, Captain Kirk, the universe is stranger than you ever imagined!

Three conceptions of biography

A biography is a narrative of a person's life. The biographer wants to tell the story of how the subject made it from childhood to adulthood; how he or she came to undertake certain actions in life; how various personal aspirations and commitments were played out in terms of extended projects with varying levels of success; how the adult's character was formed through specific experiences and influences. The biographer wants to make sense of the subject's life itinerary and character.

A biography necessarily pays attention to both external and internal factors. Externally, the biographer will investigate the nature of the family, the accidents of social class and education to which the subject was exposed, the role models and teachers by whom the subject was influenced, the habits of daily life that had unexpected consequences for the longterm development of the subject's life. Internally, the biographer will be interested in reconstructing the character, personality, and mental life of the subject; the inclinations and motives that drove the subject's choices; the values and commitments that shaped the ways in which the subject interacted with other people and social institutions.

There are several basic frames that the narrative of a biography might take. The biographer may tell a story that suggests that the subject possessed a well-developed conception of what he or she wanted out of life, and set out to take specific steps and strategies to bring about these outcomes. This can be called the "architect's view". The subject's life is understood as the fulfillment of a detailed plan, as direct as the migration pathway of a whooping crane.

At the other extreme, the biographer may find that the subject's life looks to be highly contingent and random. The subject moves from one experience to another, one opportunity and another discouragement, and the route is aimless. Call this the "random walk view." The subject's life may have had singular and valued accomplishments; but there is no overall direction to the life, only a series of stochastic and opportunistic interactions.

image: mathematical simulation of Levy walk foraging behavior (link)

There is a third possibility that incorporates both agency and contingency. The biographer may treat the subject as possessing values and ideas about the future that are guiding signposts as the subject moves through the small stages of life. The subject may have a moral sense that leads him or her to question choices through the lens of questions like "what is the right thing to do?" or "how can I contribute to a more just society?" or "what kind of person do I want to be?". Specific steps and initiatives are the result. The subject may be reflective and reflexive: he or she may consider current states of affairs and opportunities in terms of the ways in which these states of affairs contribute to half-fulfilled values and aspirations; and the subject may undertake to act in ways that give further shape to his or her mental environment -- more committed to a set of moral or spiritual values, more attentive to the satisfactions of family or creative achievement, more attuned to a sense of beauty or aesthetic balance. Call this the "bildungs view".

image: Lewis and Clark journey of discovery

The architect's view of life seems highly implausible for almost any given individual. Surely most people's lives have unfolded with a high degree of contingency, improvisation, and indeterminacy. An individual is exposed to this great teacher or that unfortunate necessity; and future actions and plans take shape partly as a result of those prior chance developments. Further, life contingency seems not to be restricted to moments of high drama, but rather seem to be distributed across the full range of daily life.

The random-walk view has some empirical plausibility as an understanding of the lives of a wide range of people. It is plausible enough to imagine that some people are highly unreflective about the future; they make decisions as necessities and opportunities come along, and they make out a life as a sum of these unrelated choices. Even the complexity introduced by the notion of the "Levy walk" is relevant to the unfolding of a life. As the diagram above indicates, the behavior of the subject is not restricted to a local domain of random choices; instead, the course is interspersed with long hops, taking the subject to a new terrain to explore randomly (link).

But it is the bildungs-view that has the most appeal for anyone who favors reflectiveness and the idea of creating a somewhat coherent whole over a long span of time. Here we are to picture a person always only partially formed, but seeking new personal and situational developments that lead towards a set of goals and values that are themselves only partially articulated. This is a boot-straps understanding of a purposive life: the individual at any given stage of life has reflected on values, goals, purposes, and satisfactions, has pursued plans and opportunities that fit with those values, and has reformulated and refined his or her values and goals as life proceeds. This is what can fairly be called a reflective life, embodying both contingency and direction.

The illustration of the 1804 Lewis and Clark expedition from St. Louis to Fort Clatsop gives a metaphorical idea of the bildungs view of a life. Lewis and Clark had some partially defined goals and plans when they set out on their journey of discovery. But much of the route depended upon contingencies encountered along the way and opportunities that were exploited in an improvisational manner. Their itinerary and changes of direction were often the result of new information, changing weather, unexpected terrain, and the like. The exact course they eventually traversed was not defined in advance; and yet there was a clear directionality to the expedition.

And what about Inspector Clouseau, depicted above in the person of Peter Sellers? The hapless inspector illustrates the random-walk model as he moves through the Pink Panther movies, with a generous amount of bumbling randomness and amazing turns of favorable luck leading to unlikely success in the end. Regrettably, no one can count on cinematic luck in ordinary life, so a bit more caution and planning seem well advised.

In framing this series of posts on rational life plans I have focused on the idea of rationality over time. There is another approach in philosophy that I haven't considered, however, that considers a similar problem from the point of view of the notion of a "meaningful life". What features or conditions make a life "meaningful" to the individual and to others? Susan Wolf's 2007 Tanner lectures provide a useful beginning at trying to analyze the good life in its fullness over time from the point of view of meaning and value (link). Also useful is Thaddeus Metz's article in the Stanford Encyclopedia of Philosophy on "the meaning of life" (link). Finally, the series of posts presented earlier on the topic of "character" is relevant to this topic as well (link). Character is relevant to life plans because it is shaped by the decisions the individual makes, and it provides the cross-temporal fiber needed to persist in a life plan under adversity.

Deliberation, rationality, and reasoning

Recent posts have raised questions about formulating a rational plan of life. This way of putting the question highlights "rationality," which has the connotation of short-term, one-off decision making. And this implication plainly does not fit the problem of life planning very well -- as noted in the two prior posts on this topic. Living a life is more like the making of a great sculpture than it is planning a Napoleonic military campaign. But what if we shifted the terms of the question and asked instead, what is involved in being deliberative and reflective about the direction of one's life? Does this give more room for bringing the idea of rationality into the idea of a life plan?

Being deliberative invokes the idea of considering one's goals reflectively and in comparison, considering strategies and actions that might serve to bring about the realization of these goals, and an ongoing consideration of the continuing validity of one's goals and strategies. Instrumental rationality takes a set of goals as being fixed; deliberative rationality works on the assumption that it is possible to reason reflectively about one’s goals themselves. This is the thrust of Socrates’ “unexamined life” — the good life requires reflection and deliberation about the things one seeks to achieve in life. Here is how Aristotle describes deliberation in the Nicomachean Ethics, book 3:

We deliberate about things that are in our power and can be done; and these are in fact what is left. For nature, necessity, and chance are thought to be causes, and also reason and everything that depends on man. Now every class of men deliberates about the things that can be done by their own efforts. And in the case of exact and self-contained sciences there is no deliberation, e.g. about the letters of the alphabet (for we have no doubt how they should be written); but the things that are brought about by our own efforts, but not always in the same way, are the things about which we deliberate, e.g. questions of medical treatment or of money-making. And we do so more in the case of the art of navigation than in that of gymnastics, inasmuch as it has been less exactly worked out, and again about other things in the same ratio, and more also in the case of the arts than in that of the sciences; for we have more doubt about the former. Deliberation is concerned with things that happen in a certain way for the most part, but in which the event is obscure, and with things in which it is indeterminate. We call in others to aid us in deliberation on important questions, distrusting ourselves as not being equal to deciding. (Nicomachean Ethics, book 3)

What Aristotle focuses on here is choice under conditions of uncertainty and complexity. Deliberation is relevant when algorithms fail -- when there is no mechanical way of calculating the absolutely best way of doing something. And this seems to fit the circumstance of planning for a life or career.

How does “deliberation” come into the question of life plans? It is essential.

(1) The goals a person pursues in life cannot be specified exogenously; rather, the individual needs to consider and reflect on his or her goals in an ongoing way. Aristotle was one of the first to reveal that often the goals and goods we pursue are, upon reflection, derivative from some more fundamental good. But Kant too had a position here, favoring autonomy over heteronomy. Reflection allows us to gain clarity about those more fundamental goods that we value.

(2) The strategies and means that we choose may have only a superficial correspondence to our goods and values which is undercut by more rigorous examination. We may find that a given mode of action, a strategy, may indeed lead to good X, but may also defeat the achievement of Y, which we also value. So deliberative reflection about the strategies and actions we choose can allow us to more fully reconcile our short-term strategies with our long-term goals and goods.

Economists and philosophers have sometimes maintained that values and goals do not admit of rational consideration. But this is plainly untrue. At the very least it is possible to discover positive and negative interactions among our goals and desires — the desire to remain healthy and the desire to eat ice cream at every meal are plainly in conflict. Less trivially, the goal of living life in a way that is respectful of the dignity of others is inconsistent with the goal of rising to power within a patriarchal or racist organization. It is possible that there are values that are both fundamental and incommensurable — so that rational deliberation and reflection cannot choose between them. But it is hard to think of examples in which this kind of incommensurability arises as a practical problem.

Consider Bruce, Jorge, and Filippo. Bruce believes that being wealthy by the age of 60 is the most important thing in his life. Jorge believes that attaining a state of spiritual fulfillment by the age of 60 is most important. And Filippo believes that having circumstances of life by age 60 in which he is involved in satisfying work, has successful family relationships and friendships, and has enough income and wealth to support a decent middle-class life is most important. Are there reasonable considerations that any one of these individuals could bring to bear against another to suggest that the other’s goals are incomplete or defective?

Aristotle addresses Bruce directly by asking how wealth could possible be a fundamental good. What does Bruce want to gain by achieving great wealth? Aren’t these activities and goods more fundamental than the wealth itself? This line of argument perhaps succeeds in persuading Bruce to give more thought to what he wants out of life — not wealth, but the things that wealth permits him to do.

Jorge may seem to be just like Bruce except he values spiritual fulfillment rather than money. Indeed, they are similar in that there is only one dimension to their life goals. But Jorge can at least maintain against Bruce that his goal is good in itself, not because of its ability to bring about some other desirable thing.

Finally, Filippo. Filippo has a more complex set of life goals, none of which reduces to a combination of the others. It is true that these goals require tradeoffs in behavior and effort; strategies that enhance friendships may depress the attainment of wealth, for example. But I think it is Filippo that we think of when we imagine a person with a reflective and deliberative life plan: a person who has identified a small but plural set of longterm goals, and who recognizes that it is necessary in the moment to find ways of balancing the attainment of one with what it takes to attain more of the other.

We might think of life planning as being less like a blueprint for action and more like a navigational guide. We might think of the problem of making intermediate life choices as being guided by a compass rather than a detailed plan — the idea that we do good work on living if we guide our actions by a set of directional signals rather than a detailed map. Life outcomes result from following a compass, not moving towards a specific GPS point on a map.

There is an analogy with business planning here. Consider the actions and plans of a CEO of a company. His or her choices in concrete decision moments are guided by several important considerations: remain profitable; prepare the ground today for viable business activity tomorrow; create an environment of trust and respect among the employees of the company; make sure that company choices also take the wellbeing of the community into account; treat employees fairly; anticipate changes in the marketplace that might dictate change in process or product within the company. But there is no certainty, no fixed prescription for success, and no algorithm for balancing the goods that the firm's leadership pursues. The successful firm will have built its success over a long series of decisions oriented towards the fundamental values of the business.

(The reference to Napoleon in Jena in the graphics above is pertinent because of the implications that Hegel drew from his experience of Napoleon as a "world historical figure". Hegel was clear that, even with a brilliant commander and a great general staff, Napoleon's ambitions in Europe were based on an unavoidably incomplete knowledge of the terrain of history. "The Owl of Minerva spreads its wings only at the falling of the dusk.")

A fresh approach to life plans

There isn't a clear philosophy of life-planning in the literature. So let's start from scratch. What do we need in order to make a plan for any temporally extended project?

• An assessment of the outcomes we want to bring about
• An assessment of the likely workings of the natural and social environment in which action will occur
• A theory about how to achieve those outcomes -- strategy and tactics
• An assessment of the likelihood of negative interactions among various aspects of the plan
• An assessment of the riskiness of the environment
• A backup plan if things go off the rails -- plan B!

We would like to arrive at a plan that has a high probability of success, and one for which there are soft landings available when future expectations are not fulfilled. If my goal is to become a symphony conductor but I also know that the qualifications needed would equally qualify me to be a performer, and if performing itself is an agreeable outcome, then aiming at conductor is less risky.

We know what it is to be rational about limited choices like choosing a new car, picking a vacation destination, or investing retirement savings. Each of these decisions falls within a broad degree of certainty of assumptions for us: we know that we enjoy the beach more than the opera, that we want a fair degree of security in our retirement accounts, or that we need a car that is good in wet weather. That is to say, we know a lot about our tastes, our future needs, and our current circumstances. So small-gauge choices like these depend fairly simply on locating a solution that serves our tastes and preferences in our current and near-future circumstances. With these conditions fixed, we can then go about the information gathering that allows us to assess how well the available sets of alternatives serve our tastes, needs, and circumstances.

Sometimes we can even reduce our choice situations to a simple set of cost-benefit tradeoffs: I'll get a 20% improvement in crash-worthiness by paying an additional $10,000 for the car I choose; I'll have a chance on a 10% annual return on an investment if I accept a greater degree of risk; etc. And I might find that I like the tradeoff for one set of costs but not for another -- more safety is worth $10,000 to me but not $50,000. Or I will accept the greater investment risk when it means moving from 1% chance of losing everything to a 3% chance, but not to a 10% chance.

A life plan isn't like this, however. Consider the space of choices that confronts the 20-year old college student Miguel: what kind of work will satisfy me over the long term? How much importance will I attribute to higher income in twenty years? Do I want to have a spouse and children? How much time do I want to devote to family? Do I want to live in a city or the countryside? How important to me is integrity and consistency with my own values over time? These kinds of questions are difficult to answer in part because they don't yet have answers. Miguel will become a person with a set of important values and commitments; but right now he is somewhat plastic. It is possible for him to change his preferences, tastes, values, and concerns over time. So perhaps his plan needs to take these kinds of interventions into account.

Another source of uncertainty has to do with the future of the world itself. Will the economy continue to provide decent opportunities for young people, or will income stratification continue to increase? Will climate change make some parts of the world much more difficult for survival? Will religious strife worsen so that safety is very difficult to achieve? Is Mary Poppins or William Gibson the better prognosticator of what the world will look like in thirty years? A plan that looks good in a Mary Poppins world may look much worse in the Sprawl (Gibson's anti-utopian city of the future).

And then there is the difficult question of akrasia -- weakness of the will. Can I successfully carry out my long term plans? Or will short term temptations make it impossible for me to sustain the discipline required to achieve my long term goals? (Somewhere Jon Elster looks at this problem as a collective action problem across stages of the self. Is this a reasonable approach?) For that matter, how much should future goods matter to me in the present?

It is worth asking whether life plans actually exist for anyone. Perhaps most people's lives take shape in a more contingent and event-driven way. Perhaps guided opportunism is the best we are likely to do: look at available opportunities at a given moment, pursue the opportunity that seems best or most pleasing at that point, and enjoy the journey. Or perhaps there are some higher-level directional rules of thumb -- "choose current options that will contribute in the long run to a higher level of X". In this scenario there is no overriding plan, just a series of local choices. This alternative is pretty convincing as a way of thinking about the full duration of a person's life, as any biographer is likely to attest.

Consider an analogy with the life of a city or state: decisions and policies are established at various points in time. These decisions contribute to the life course of the city; monuments established in 200 BCE continued to inform Roman life in 300 AD. But Rome was indeed not built in a day, and its eventual course was not envisaged or planned by any of its founders and leaders. A city's "life" is the complex resultant of deliberation at many points in time, struggle, and contingency. And perhaps this describes a person's life as well.

This point of view has a lot in common with Herbert Simon's 1957 concept of bounded rationality and satisficing rather than maximizing as a rule of rational decision-making (Models of Man). Instead of heroically attempting to plan for all contingencies over the full of one's life, a bounded approach would be to consider short periods and make choices over the opportunity sets available during those periods. And if we superimpose on these choices a higher-level set of goals to be achieved -- having time with family, living in conformity to one's moral or religious values, gaining a set of desired character traits -- then we might argue that this decision-making process will be biased towards outcomes that favor one's deeper values as well as one's short-term needs and interests.

This approach will not optimize choices over the full lifetime; but it may be the only approach that is feasible given the costs of information gathering and scenario assessment.

So what about a rational life plan? At this point the phrase seems inapropos to the situation of a person's relationship to his or her longterm "life". A life is more of a concatenation of a series of experiences, projects, accidents, contingencies -- not a planned artifact or painting or building. A life is not a novel, a television series, or a mural with an underlying storyboard in which each element has its place. And therefore it seems inapt to ask for a rational plan of life. Individuals make situated and bounded deliberative decisions about specific issues. But they don't plot out their lives in detail. 

What seems more credible is to ask for a framework of navigation, a set of compass points, and a general set of values and purposes which get invested through projects and activities. The idea of the bildungsroman seems more illuminating -- the idea of a young person taking shape through a series of challenging undertakings over time. Development, formation, values clarification, and the formation of character seem more true to what we might like to see in a good life than achieving a particular set of outcomes.

Where, then, do thinking and reasoning come into the picture? This is where Socrates and Montaigne seem to be relevant. They look at living as an opportunity for deepening self-knowledge and articulation of values and character. "To philosophize is to learn how to die" (Montaigne) and "The unexamined life is not worth living" (Socrates). The upshot of these aphorisms seems to be this: reasoning and philosophizing allow us to probe, question, and extend our values and the things we strive for. And having examined and probed, we are also in a position to assess and judge the actions and goals that are presented to us at various stages of life. How does a college major, a first job, a marriage, or a parenting challenge frame the future into which the young person develops? And how can practical reflection about one's current values help to give direction to the future choices he or she makes later in life? 

Practical rationality perhaps amounts to little more than this when it comes to constructing a life: to consider one's best understanding of the goods he or she cares most about, and acting in the present in ways that shape the journey towards a future that better embodies those goods for the person and his or her concerns.

How to do cephalopod philosophy

How should researchers attempt to investigate non-human intelligence? The image above raises difficult questions. The octopus is manipulating (tenticlating?) the Rubik's cube. But there are a raft of questions that are difficult to resolve on the basis of simple inductive observation. And some of those questions are as much conceptual as they are empirical. Is the octopus "attempting to solve the cube"? Does it understand the goal of the puzzle? Does it have a mental representation of a problem which it is undertaking to solve? Does it have temporally extended intentionality? How does octopus consciousness compare to human consciousness? (Here is a nice website by several biologists at Reed College on the subject of octopus cognition; link.)

An octopus-consciousness theorist might offer a few hypotheses:

1. The organism possesses a cognitive representation of its environment (including the object we refer to as "Rubik's cube").
2. The organism possesses curiosity -- a behavioral disposition to manipulate the environment and observe the effects of manipulation.
3. The organism has a cognitive framework encompassing the idea of cause and effect.
4. The organism has desires and intentions.
5. The organism has beliefs about the environment.
6. The organism is conscious of itself within the environment.

How would any of these hypotheses be evaluated?

One resource that the cephalopod behavior theorist has is the ability to observe octopi in their ordinary life environments and in laboratory conditions. These observations constitute a rich body of data about behavioral capacities and dispositions. For example:

Here we seem to see the organism conveying a tool (coconut shell) to be used for an important purpose later (concealment) (link). This behavior seems to imply several cognitive states: recognition of the physical characteristics of the shell; recognition of the utility those characteristics may have in another setting; and a plan for concealment. The behavior also seems to imply a capacity for learning -- adapting behavior by incorporating knowledge learned at an earlier time.

Another tool available to the cephalopod theorist is controlled experimentation. It is possible to test the perceptual, cognitive, and motor capacities of the organism by designing simple experimental setups inviting various kinds of behavior. The researcher can ask "what-if" questions and frame experiments that serve to answer them -- for example, what if the organism is separated from the shell but it remains in view; will the organism reaquire the shell?

A third tool available to the cephalopod researcher is the accumulated neuro-physiology that is available for the species. How does the perceptual system work? What can we determine about the cognitive system embodied in the organism's central nervous system?

Finally, the researcher might consult with philosophers working on the mind-body problem for human beings, to canvass whether there are useful frameworks in that discipline that might contribute to octopus-mind-body studies. (Thomas Nagel's famous article, "What is it Like to Be a Bat?", comes to mind, in which he walks through the difficulty of imagining the consciousness of a bat whose sensory world depends on echo-location; link.)

In short, it seems that cephalopod cognition is a research field that necessarily combines detailed empirical research with conceptual and theoretical framing; and the latter efforts require as much rigor as the former.

Rationality over the long term

image: Dietrich Bonhoeffer with his students

Millions of words have been written on the topic of rationality in action. Life involves choices. How should we choose between available alternatives? Where should I go to college? Which job should I accept? Should I buy a house or rent an apartment? How much time should I give my job in preference to my family? We would like to have reasons for choosing A over B; we would like to approach these choices "rationally."

These are all "one-off" choices, and rational choice theory has something like a formula to offer for the decider: gain the best knowledge available about the several courses of action; evaluate the costs, risks, and rewards of each alternative; and choose that alternative that produces the greatest expected level of satisfaction of your preferences. There are nuances to be decided, of course: should we go for "greatest expected utility" or should we protect against unlikely but terrible outcomes by using a maximin rule for deciding?

There are several deficiencies in this story. Most obviously, few of us actually go through the kinds of calculations specified here. We often act out of habit or semi-articulated rules of thumb. Moreover, we are often concerned about factors that don't fit into the "preferences and beliefs" framework, like moral commitments, conceptions of ourselves, loyalties to others, and the like. Pragmatists would add that much mundane action flows from a combination of habit and creativity rather than formal calculation of costs and benefits.

But my concern here is larger. What is involved in being deliberative and purposive about extended stretches of time? How do we lay out the guideposts of a life plan? And what is involved in acting deliberatively and purposively in carrying out one's life plan or other medium- and long-term goals?

Here I want to look more closely than usual at what is involved in reflecting on one's purposes and values, formulating a plan for the medium or long term, and acting in the short term in ways that further the big plan. My topic is "rationality in action", but I want to pay attention to the issues associated with large, extended purposes -- not bounded decisions like buying a house, making a financial investment, or choosing a college. I'm thinking of larger subjects for deliberation -- for example, conquering all of Europe (Napoleon), leading the United States through a war for the Union ( Lincoln), or becoming a committed and active anti-Nazi (Bonhoeffer).

The scale I'm focusing on here corresponds to questions like these:

• How did Napoleon deliberate about his ambitions in 1789? How did he carry out his thoughts, goals, and plans?
• How did Abraham Lincoln think about slavery and the Union in 1861? How did his conduct of politics and war take shape in relation to his long term goals?
• How did Richard Rorty plan his career in the early years? How did his choices reflect those plans? (Neil Gross considers this question in Richard Rorty: The Making of an American Philosopher; link.)
• How did Dietrich Bonhoeffer deliberate about the choices in front of him in Germany in 1933? How did he decide to become an engaged anti-Nazi, at the eventual cost of his life?

What these examples have in common is large temporal scope; substantial uncertainties about the future; and extensive intertwining of moral and political values with more immediate concerns of self-interest, prudence, and desire. Moreover, the act of formulating plans on this scale and living them out is formative: we become different persons through these efforts.

The intriguing question for me at the moment is the issue of rational deliberation: to what extent and through what processes can individuals engage in a rational process in thinking through their decisions and plans at this level? Is it an expectation of rationality that an individual will have composed nested sets of plans and objectives, from the most global to the intermediate to the local?

Or instead, does a person's journey through large events take its shape in a more stochastic way: opportunities, short term decisions, chance involvements, and some ongoing efforts to make sense of it all in the form of a developing narrative? Here we might say that life is not planned, but rather built like Neurath's raft with materials at hand; and that rationality and deliberation come in only at a more local scale.

Here is a simple way of characterizing purposive action over a long and complex period. The actor has certain guiding goals he or she is trying to advance. It is possible to reflect upon these goals in depth and to consider their compatibility with other important considerations. This might be called "goal deliberation". These goals and values serve as the guiding landmarks for the journey -- "keep moving towards the tallest mountain on the horizon". The actor surveys the medium-term environment for actions that are available to him or her, and the changes in the environment that may be looming in that period. And he or she composes a plan for these circumstances-- "attempt to keep moderate Southern leaders from supporting cecession". This is the stage of formulation of mid-range strategies and tactics, designed to move the overall purposes forward. Finally, like Odysseus, the actor seizes unforeseen opportunities of the moment in ways that appear to advance the cause even lacking a blueprint for how to proceed.

We might describe this process as one that involves local action-rationality guided by medium term strategies and oriented towards long term objectives. Rationality comes into the story at several points: assessing cause and effect, weighing the importance of various long term goals, deliberating across conflicting goals and values, working out the consequences of one scenario or another, etc.

As biologists from Darwin to Dawkins have recognized, the process of species evolution through natural selection is inherently myopic. Long term intelligent action is not so, in that it is possible for intelligent actors to consider distant solutions that are potentially achievable through orchestrated series of actions -- plans and strategies. But in order to achieve the benefits of intelligent longterm action, it is necessary to be intelligent at every stage -- formulate good and appropriate distant goals, carefully assess the terrain of action to determine as well as possible what pathways exist to move toward those goals, and act in the moment in ways that are both intelligent solutions to immediate opportunities and obstacles, and have the discipline to forego short term gain in order to stay on the path to the long term goal. But, paradoxically, it may be possible to be locally rational at every step and yet globally irrational, in the sense that the series of rational choices lead to an outcome widely divergent from the overriding goals one has selected.

I've invoked a number of different ideas here, all contributing to the notion of rational action over an extended time: deliberation, purposiveness, reflection, calculation of consequences, intelligent problem solving, and rational choice among discrete alternatives. What is interesting to me is that each these activities is plainly relevant to the task of "rational action"; and yet none reduces to the other. In particular, rational choice theory cannot be construed as a general and complete answer to the question, "what is involved in acting rationally over the long term?".

Michael Bratman is the philosopher who has thought about these issues the most deeply; Intention, Plans, and Practical Reason. Manuel Vargas and Gideon Yaffe's recent festschrift on Bratman's work, Rational and Social Agency: The Philosophy of Michael Bratman, is also a useful contribution on the subject. Sarah Paul provides a nice review of Rational and Social Agency here.

Amartya Sen's commitments

A recent post examined the Akerlof and Kranton formalization of identity within a rational choice framework.  It is worth considering how this approach compares with Amartya Sen's arguments about "commitments" in "Rational Fools" (link).

Sen's essay is a critique of the theory of narrow economic rationality to the extent that it is thought to realistically describe real human deliberative decision-making. He chooses Edgeworth as a clear expositor of the narrow theory: "the first principle of Economics is that every agent is actuated only by self interest" (Sen 317, quoting Mathematical Psychics). Sen notes that real choices don't reflect the maximizing logic associated with rational choice theory: "Choice may reflect a compromise among a variety of considerations of which personal welfare may be just one" (324). Here he argues for the importance of "commitments" in our deliberations about reasons for action. Acting on the basis of commitment is choosing to do something that leads to an outcome that we don't subjectively prefer; it is acting in a way that reflects the fact that our actions are not solely driven by egoistic choice.  "Commitments" are other-regarding considerations that come into the choices that individuals make.

Sen distinguishes between sympathy and commitment:

The former corresponds to the case in which the concern for others directly affects one's own welfare. If the knowledge of torture of others makes you sick, it is a case of sympathy; if it does not make you feel personally worse off, but you think it is wrong and you are ready to do something to stop it, it is a case of commitment. (326)

The characteristic of commitment with which I am most concerned here is the fact that it drives a wedge between personal choice and personal welfare, and much of traditional economic theory relies on the identity of the two. (329)

Sen thinks that John Harsanyi made an advance on the narrow conception of rationality by introducing discussion of two separate preference orderings that are motivational for real decision-makers: ethical preferences and subjective preferences. (This is in "Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparisons of Utility".)  But Sen rightly points out that this construction doesn't give us a basis for choosing when the two orderings dictate incompatible choices.  Sen attempts to formalize the idea of a commitment as a second-order preference ordering: a ranking of rankings.  "We need to consider rankings of preference rankings to express our moral judgments" (337).

Can one preference ordering do all these things? A person thus described may be "rational" in the limited sense of revealing no inconsistencies in his choice behavior, but if he has no use for these distinctions between quite different concepts, he must be a bit of a fool. The purely economic man is indeed close to being a social moron. Economic theory has been much preoccupied with this rational fool decked in the glory of his one all-purpose preference ordering. To make room for the different concepts related to his behavior we need a more elaborate structure. (335-336)

Here is an example.  "I wish I liked vegetarian foods more" is an example of a second-order preference ranking: it indicates a rational preference for the first-order ranking in which the vegetarian option comes ahead of the lamb option over the ranking in which these options are reversed.  And Sen's point is an important one: the second-order ranking can be behaviorally influential.  I may choose the vegetarian option, not because I prefer it, but because I prefer the world arrangement in which I go for the vegetarian option.  Or in other words, one's principles or commitments may trump one's first-order preferences.

Significantly, Sen's thinking on this subject was developed in part through a conference organized by Stephen Körner on practical reason in the 1970s (Practical Reason: Papers and Discussions).  This is significant because it focuses attention on a very basic fact: we don't yet have good theories of how a variety of considerations -- ethical principles, personal identities, feelings of solidarity, reasoning about fairness, and self-interest -- get aggregated into decisions in particular choice circumstances.

Other economists might object to this formulation on the basis of the fact that second-order preference rankings are more difficult to model; so we don't get clean, simple mathematical representations of behavior if we introduce this complication.  Sen acknowledges this point:

Admitting behavior based on commitment would, of course have far- reaching consequences on the nature of many economic models. I have tried to show why this change is necessary and why the consequences may well be serious. Many issues remain unresolved, including the empirical importance of commitment as a part of behavior, which would vary, as I have argued, from field to field. I have also indicated why the empirical evidence for this cannot be sought in the mere observation of actual choices, and must involve other sources of information, including introspection and discussion. (341-342)

But his reply is convincing.  There are substantial parts of ordinary human activity that don't make sense if we think of rationality as egoistic maximization of utility.  Collective action, group mobilization, religious sacrifice, telling the truth, and working to the fullest extent of one's capabilities are all examples of activity where narrow egoistic rationality would dictate different choices than those ordinary individuals are observed to make.  And yet ordinary individuals are not irrational when they behave this way. Rather, they are reflective and deliberative, and they have reasons for their actions.  So the theory of rationality needs to have a way of representing this non-egoistic reasonableness.  This isn't the only way that moral and normative commitments can be incorporated into a theory of rational deliberation; but it is one substantive attempt to do so, and is more satisfactory (for me, anyway) than the construction offered by Akerlof and Kranton.

(I also like the neo-Kantian approach taken by Tom Nagel in The Possibility of Altruism as an effort to demonstrate that non-egoistic reasoning is rational.)