Todd and Gigerenzer, “Précis of Simple Heuristics That Make Us Smart” – (Olivia Brigstocke and Nosagie Asaolu)

We will summarize the text,  offer our perspective on the author’s’ main points and pose some questions.

Unbounded Rationality

Todd and Gigerenzer propose two types of rationality: Bounded Rationality and Demonic (Divine) Reasoning. The latter assumes that “the human mind has essentially unlimited demonic or supernatural reasoning power” and is divided further into:

  1. Unbounded rationality
  2. Optimization under constraints.

Unbounded rationality represents a system that is omniscient and capable of certainty with unlimited access to information, however such a model is not realistic as the human mind has limited access to knowledge, time, money, and attention, when making decision.

The constrained optimization model requires a “stopping point” in the otherwise endless pursuit of knowledge. This point is reached when the costs of continuing to search for information outweigh the benefits of the additional knowledge that may be gained. This process also assumes almost unlimited time and computational power, as calculating the potential cost and benefit of each piece of further information would be hugely expensive and ultimately impossible.

The authors argue that “reasoning can be powerful and accurate without requiring unlimited time and knowledge”, which is why they move from the demonic model to the bounded model. Many philosophers and economists, however, argue that bounded rationality models “use a greater number of parameters and become more demanding mathematically” (730) than their unbounded alternatives, and for this reason they choose to stick with unbounded rationality. Todd and Gigerenzer respond to these critics by saying that “rationality need not be optimization, and bounds need not be constraints”(730).

Bounded Rationality

Thus, they present a notion of “Bounded Rationality” which presents the process of decision making as that of “a bounded [human] mind reaching into an adaptive toolbox filled with fast and frugal heuristics”(729). Bounded rationality accounts for both “the limitation of the human mind” and “the structure of the environment in which the mind operates”. The latter is a central theme of “Ecological rationality”, which shows the extent to which a heuristic is adapted to the structure of its operating environment. Subsequently, the authors present notions of “Satisficing” and “Fast and Frugal Heuristics”.

Satisficing is a method of making a choice from a set of options encountered sequentially when one “does not know much about the possibilities in advance”. It utilizes an “aspiration level” and stops searching for alternatives once an option that exceeds the aspiration level is found. However, difficulties arise during both the computation of the aspiration level and its relationship with each encountered option.

Fast and Frugal Heuristics limit the search of objects by using “easily computable stopping rules”, and make choices with “easily computable decision rules”. This represents Bounded Rationality in its purest form insofar as some fast and frugal heuristics make “satisficing sequential option decisions” and some do not. An example of a fast and frugal heuristic is One Reason decision making where one single piece of information is used to make a choice. This sharply contrasts with heuristics of unbounded rationality which attempt to assimilate unlimited information in different decision making environments.

The authors explain the fast and frugal heuristic principles that:

  1. Guide search
  2. Stop search
  3. Make decisions based on (1) and (2)

Regarding Search guidance, fast and frugal heuristics use ‘simple search strategies’, such as cue orderings, which are easy to compute and ignore dependencies between cues.  This contrasts with ‘active search strategies’ which use extensive computations or knowledge to find the next cue or alternative to search.

Regarding Stopping search, fast and frugal heuristics employ simple rules. For example: cease searching and make a decision as soon as a reason that favors one alternative is found. For alternatives, simple aspiration level stopping rules can be used.

Regarding principles for decision making(3), decisions can be based on only one reason, regardless of the total number of cues found during search. Also, cues are not weighed or combined into a “common currency”. Rather, a simple elimination process is employed until one final choice remains.

The authors note that the components of fast and frugal heuristics are constructed by combining building blocks, such as One Decision making, or nesting existing heuristics within each other, such as a Recognition Heuristic that works on the basis of an elementary cognitive capacity and recognition memory.

Fast and frugal heuristics are divided into four classes, which are :

  1. Ignorance Based Decision making
  2. One Reason Decision making
  3. Elimination Heuristics for Multiple-option choices
  4. Satisficing Heuristics

 

First, Ignorance based decision making ensures that, in two-option situations, decision makers choose recognized options over unrecognized options. The authors give an example in which humans are more likely to choose recognizable ham over “odd colored” eggs. Such ignorance based recognizing heuristics perform better than random choice in environments where a positive correlation exists between the decision maker’s exposure to different possibilities and the ranking of these possibilities along the decision criteria being used. In other words, the more recognizable an option is, the more likely it is that it will get chosen.

Second, One Reason decision making ensures that, when multiple cues exist, a stopping rule is activated once enough information has been gathered to make a decision. In other words, the decision maker stops looking for cues as soon as a cue is found that differentiates between the options considered. Hence, such heuristics include both a stopping rule, and a decision cue. Moreover, cues can be searched by the following methods:

  1. i) In order of validity (Take the Best heuristic)
  2. ii) According to the order determined by their past successes in stopping search

iii) In random order (Minimalist heuristic)

The authors note that such heuristics are non-compensatory as “once they have used a simple cue to make a decision, no further cues in any combination can undo or compensate for that one cue’s effect”.

Third, Elimination Heuristics for multiple option choices are employed when each available cue “dimension” has fewer values than the number of available alternatives. One Reason decision making is insufficient in such cases because a single cue cannot distinguish between all alternatives. For example, knowing whether or not 15 cities contain a river is not enough to decide which is the most habitable. Moreover, elimination heuristics employ both stopping rules and elimination. The former ensures that cues are only sought until enough is known to make a decision while the latter ensures that successive cues are used to eliminate options until a single option remains.

Fourth, unlike other heuristics that assume all possible options are available to the decision maker, Satisficing heuristics find alternatives sequentially over an extended period or spatial region. This is done by setting an aspiration level and the search for alternatives is stopped once the selection criterion is met.

The authors posit that organisms deal with their environments. Hence, different environments can have different specific fast and frugal heuristics that exploit the information structures present to make adaptive decisions. However, this could lead to a multitude of specific fast and frugal heuristics. Thus, the authors propose the following information structures which make fast and frugal heuristics “ecologically rational”:

  1. Non-compensatory Information
  2. Scarce Information
  3. J-shaped distributions
  4. Decreasing Populations

Regarding Non-compensatory information, the potential contribution of each new cue falls rapidly. In such cases, the Take the Best heuristic is ecologically rational.

Regarding Scarce information, the Take the Best heuristic outperforms  a class of linear models on average when few cues are known relative to the number of objects.

In J-Shaped distributions there are many small values and few large values. In such cases, the QuickEst heuristic, which takes values of objects along the same criterion using as little information as possible, is ecologically rational.

Regarding decreasing populations where the set of options is constantly shrinking, a satisficing heuristic is ecologically rational.

 

Furthermore, the authors propose that fast and frugal heuristics should be evaluated based on correspondence criteria regarding speed, frugality and accuracy. This contrasts with coherence criteria which are primarily concerned with the internal logical coherence of judgements. Moreover, this allows for the comparison of heuristics within the actual requirements of its environment which may include, using a minimal amount of information or making accurate decisions in a minimal amount of time.

 

Lastly, the the authors conceive of the mind as an “adaptive toolbox” which is a collection of specialized cognitive mechanisms that evolution has built into the human mind for specific domains of inference and reasoning. These contain psychological adaptations and include perceptual and memory processes. However, higher order cognitive processes are better modeled by simpler algorithms than lower order mechanisms.

Conclusion

The authors conclude by stating the following challenges to be addressed by their research:-

  1. Cognitive tasks such as planning, perceptual mechanisms, and problem solving
  2. Adaptive problems regarding domain specificity and the organization of the mind
  3. Social norms and emotions
  4. Ecological irrationality regarding the aspects that shape the design and performance of decision heuristics
  5. Performance criteria, that is, is there a role for coherence criteria?
  6. Selecting heuristics, in other words, how does the mind know which heuristic to use?
  7. Multiple methodologies regarding the prevalence of simple heuristics in human and animal reasoning

The authors’ conception of fast and frugal heuristics, although probable, seems to be lacking in both empirical evidence and in the complexity of the human decision making process. Regarding the former, it is acknowledged that empirical means of testing search processes during human decision making are either non-existent or not utilisable. This is problematic as it leaves one wondering what theory of knowledge and categorization of epistemology fast and frugal heuristics adhere to. Regarding the latter, they seem to overlook the complexity of nesting heuristics, calculating the validity of cues, and the distinction between conscious and unconscious rationality, if existent. Lastly, one is forced to ask whether the authors are trying to understand cognition in the context of modeling rational systems, and/or understand how humans reason. In other words, do humans inherently embody ecological rationality in ways that artificial systems cannot (example: early vision and visual recognition).

Questions

  1. Does this paper serve as an argument in support of the naturalisation of epistemology? If so, is it sufficiently supported by empirical findings? If not, is it supported by an externalist, coherence, or foundationalist theory of knowledge?
  2. At what point does the nesting of different fast and frugal heuristics require extensive computation, time, and knowledge? Think about making simple vs complex decisions
  3. How do we measure the validity of a cue in relation to other cues? Do Todd and Gigerenzer oversimplify the cue selection process?

9 thoughts on “Todd and Gigerenzer, “Précis of Simple Heuristics That Make Us Smart” – (Olivia Brigstocke and Nosagie Asaolu)”

  1. On p. 729 the authors say “The greatest weakness of unbounded rationality is that it does not describe the way real people think.” This may be true, but from a philosophical perspective, does unbounded rationality give us an ideal normative standard, a model of the way we “should” think?

    I found the idea of an “adaptive toolbox” compelling (p. 740). Do others agree that we may have a variety of “serviceable solutions” to choose from when solving problems or making decisions, and that we are not limited to a single cognitive mechanism? Or does allowing for the functioning of different strategies in different environments weaken their case for fast and frugal heuristics?

  2. As mentioned by previous commenters, I also started making comparisons to ECHO’s model of reasoning when reading this article. To me, Todd and Gigerenzer’s “fast and frugal” heuristics model serves as a better representation or approximation of human reasoning by working in the limits of human rationality. While ECHO’s aim is to arrive at the correct conclusion based on optimal reasoning, “fast and frugal” heuristics are used to arrive at the best conclusion based on a limited bank of knowledge. The former, while useful in determining an outcome free from human computational limits and biases, cannot be used to model or as a replacement for human reasoning. Heuristics better model this, but – as many artificial models that attempt to mimic the natural world do – fails to correctly mimic human reasoning.

    Heuristics are an evolutionarily advantageous response to the “bounded” nature of human knowledge. We cannot know everything and we most certainly cannot call the entirety of our knowledge all at once. An important aspect of the heuristic method of reasoning is the ability to “nest” heuristics, that is to use more general beliefs to rule out conclusions about specific beliefs and vice versa. This is an act of inhibition, which is critical at all levels of cognition.

    This relationship does not support the naturalization of epistemology, as the heuristic model cannot really be tested empirically. This is a case in which a traditional epistemological approach may be better equipped to answer a question by drawing from natural science theories (in this case, psychological theories of cognition and the theory of evolution) to provide premises for a conclusion.

  3. In this article, Todd and Gigerenzer propose another model of human rationality. As I was reading the article, all of my initial reactions were a result of comparing this model system to ECHO. While I struggled to come to terms with the simplicity of ECHO, this model system poses even more problems for me. Most notably, the one-reason decision-making model described on pages 733-735 is far too simplistic. Although the “Take the Best” method (in which the decision is based upon the most-often correct factor) proved to have a high accuracy in decision-making, I believe there is something fundamentally incorrect with this process. While I certainly believe that humans are limited in their ability to make decisions due to time, access to knowledge, and mental capacities, many facets of one’s beliefs, emotions, and knowledge banks are important in decision making, and excluding most of these for the sake of frugality and speed seems to be innately vulnerable to incorrect decisions. While we spoke about how ECHO might actually be extremely useful in scientific and even legal settings, I find it hard to believe that one-reason decision making would be accurate in these environments. In science, more often than not, confounding and unexpected variables play a critical role in experimental outcome. For example, in clinical trials, alleviation of symptoms may be the most accurate predictor of effectiveness of a therapeutic. However, what about side effects? What about gender and age differences in response? These need to be taken into account in order to weigh costs and benefits. Even the architecture of trees in the general “fast and frugal” heuristic model is problematic. In the medical setting used as an example at the beginning of the article (page 727), while this simple decision tree may work in most cases, other diseases present in the patient may completely change that tree, indicating that the tree is a gross underestimation of the power of experience with prior patients and the medical school education. Therefore, knowledge of the relevant environment and premises is critical, and should not be ignored. That being said, Todd and Gigerenzer provide examples in which these methods have worked. Are there any general environment in which this model may be useful?

    On page 731, Todd and Gigerenzer provide a reason for using fast and frugal heuristics: “combining information from different cues requires converting them into common currency, a conversion that may be expensive if not actually impossible.” Isn’t this what we do in our every day life? Do we not take what we believe and how we feel and pit those factors against each other to make a decision?

    Finally, assuming that using these tree models is an acceptable way of making decisions, as Marisa asks, how do we construct these trees? What computation models do we use? This seems like a meta-question to me, as this questions asks: how do we make decisions about how to make decisions?

  4. I agree with the authors in that “there is a point where too much information and too much information processing can hurt” (737). However, I think that Todd and Gigerenzer’s use of simple heuristics is only applicable to a specific type of decision-making scenario. For instance, it could beneficial to use a simple heuristics model in a fast paced scenario like a hospital, in which physiological conditions can be categorize in a set range of danger, and thus prioritize medical emergencies. However, using such model on a hospital setting can be seen as unmoral or even inhumane since an actual machine will determine the fate of a human being. One could argue that a doctor would try to do everything they can to try to safe the person, even if the changes are low, whereas three yes/no cues perhaps will not prioritize a person (727). Furthermore, I found Todd and Gigerenzer’s point about recognition heuristic bit problematic. The authors explain “entire fast and frugal heuristics can themselves be combined by nesting one inside the other” (732), which makes me wonder if is quality is being sacrifice in exchange for simplicity (speed and frugality). For example, I would agree with Deniz that Dr. Seuss’s example shows a lack of expansion of knowledge. Todd and Gigerenzer explain that “things that we do not recognize in our environment are more often than not inedible, because humans have done a reasonable job of discovering and incorporating edible substances into our diet” (732). However, our life-long environments consist of more than environments we recognized. Humans are exposed to different type of environmental factors and if simple heuristics encourage us to choose only what we recognized through past experiences- or as Deniz put it, “to play it safe”- then I don’t see how simple heuristics can benefit the adaptiveness of human rationale. Additionally, like Todd and Gigerenzer described, “there are no optimal strategies in many real-world environments” (738). However, I also wonder if they are simplifying decision-making processes. Also, thinking of their cost-benefit description, I wonder if simple heuristics is only applicable immediate cost-benefit scenarios.

  5. Todd and Gigerenzer open their paper with one example of a decision tree used to assess a patient that has been rushed into a hospital. The authors note that “while decision trees are generally easy to use, their construction in the first place can be computationally expensive” (728). How do we construct these decision trees?

    Todd and Gigerenzer divide rationality into different forms (Figure 2, 729). Like bounded reality, optimization under constraints involves “limited information search” and was originally built as a more realistic model that respects the limitations of human minds. However, this approach models “limited” search by “assuming the mind has essentially unlimited time and knowledge with which to evaluate the costs and benefits of further information search” (730). Todd and Gigerenzer note that some philosophers argue real humans have the ability to perform massive computations required for constrained optimization. Do artificial systems have the ability to perform these computations?

    Bounded rationality seems to better explain the heuristics that “capture how real minds make decisions under constraints of limited time and knowledge” (728). Furthermore, unbounded rationality doesn’t describe the way people think because the human mind has these constraints. Why do philosophers argue that bounded rationality models “use a greater number of parameters and become more demanding mathematically” than their unbounded alternatives? (730).

  6. I am a lot less skeptical about the use of simple heuristics and more about the context in which it is used. Todd & Gigerenzer are very clear about the idea that heuristics are not an attempt at optimal rational decision-making because they believe there is no such thing given the restraints of a real-world environment. They are simply proposing that heuristics are a “faster, more frugal, and more accurate” when it comes to decision tasks (738). I am a lot more inclined to trust this model in a demanding, fast paced decision-making environment (such as a hospital) where perhaps emotional decisions hinder the best possible outcomes. Putting professional experience and intuition gleaned over a prolonged career above a simple heuristic algorithm seems problematic only because it devalues the education and experiences that put a doctor in their position, reducing their role from a decision-maker to a direction-follower—a role not predicated on a once valuable medical-school degree. The lack of emotional or subjective experiences indicates to me that this model seeks to be less descriptive and heavily prescriptive. If their models truly lead to the most accurate/desirable outcomes in the shortest amount of time then their model should be trusted as a means of decision-making in a time-dependent environment. My skepticism however, comes from how the decision-making fits together with a model of justification as well as other contexts in which the fastest decision may not lead to the best justification. For example, in explaining and justifying a belief such as evolution, we are not operating under a demanding environmental pressure to reach a reach a sound justification. I would rather have a sound, logical collection of evidence to ultimately lead in the justifying of such beliefs. Todd & Gigerenzer attempt to address this issue they call “over-fitting” an environment by “using only one or a few of the most useful cues” (737). This still doesn’t address a scenario in justifying a complex belief versus making a most-justified decision. Such examples are the point in which “nesting of different fast and frugal heuristics requires extensive computation, time, and knowledge.” Since complex beliefs are excluded from Todd and Gigerenzer’s model, have simple heuristics offered a plausible model in reaching our goal of what best justifies a belief, or have they only offered a piece of the puzzle?

  7. I agree that the recognition heuristic is, for the most part, adaptive (732). Of course one would choose the familiar option if hoping to avoid dangerous and uncertain circumstances. Hopefully I’m not being trivial or nitpicky here, but I don’t think Todd and Gigerenzer offer a great example in their illustration of this point. Dr. Seuss’ menu choice of green eggs or ham (732) is not exactly a life-threatening scenario. (The authors note that what’s at stake is only the quality of the meal.) Really, the only thing the recognition heuristic does in this case is help us “play it safe.” But if we’re always playing it safe, when it wouldn’t be too problematic if we didn’t (i.e., if we tried something weird on the menu, it wouldn’t really be all that bad), how do we expand our knowledge of what’s edible and good so that our recognition heuristic is continually updated and enhanced? That is all to say, we have to break the rules occasionally, if we wish to improve our heuristics. Does Todd and Gigerenzer’s research program allow for such rule-breaking?

  8. Although I agree that the authors’ argument for simple heuristics is lacking in empirical evidence and complexity, I appreciated that they discussed “ecological rationality” and the way evolution has influenced human reasoning. While other models we have looked at have largely, “endowed organisms with unlimited abilities to know, memorize, and compute”, I found the idea of simple heuristics a much more realistic and plausible model of human reasoning (at least in the situations that the author described). I could not help also questioning, as Bia and Nosagie did, where the author would stand on the Replacement Thesis. It is clear that the authors believe that the philosophical focus on the “coherence” in research on reasoning and decision making has detracted from our study of human judgement and “…how it fares in real-world environments” (737), but they admit that the data they have been able to collect have been highly variable and that their “…successes have been modest in the face of the challenges that remain” (740) in studying the use of simple heuristics. I assume conducting studies of human reasoning in this way falls under naturalized epistemology because it acknowledges the great impact of the environment and evolution, but do these insights necessarily render traditional epistemology “meaningless”? Or would the authors concede that there are still holes that remain that traditional epistemology must “fill in”, and these authors would advocate for a transformative epistemology?
    Second, I found the authors use of correspondence criteria rather than coherence criteria appealing, at least to an extent. The way I understood it, coherence is primarily concerned with internal “laws of logic and probability” while correspondence is concerned with how judgements help us make useful decisions within our external environment (737). The authors compare the mind and environment as a couple who have “become estranged” as a result of the philosophical focus on coherence in studying reasoning and decision and argue that correspondence criteria is superior and should replace multiple coherence criteria (738). In my opinion, it seems logical that both the internal and external worlds play a role in human cognition. Our judgements are affected by internal logic but also surely by our environment as well. Why do the authors wish to pick one criteria over the other? Could they both exist together in some way?

  9. Indubitably, Todd and Gigerenzer oversimplify the cue selection process. Though the process is more complex than is primarily observed, they write, “Fast and frugal search-guiding principles do not use extensive computations or knowledge to determine where to look next. But such simplicity need not lead to a disadvantage in decision accuracy, because simple search strategies can help heuristics to be more robust than those that attempt to optimize their information search” (731). They mock computational models, but these models result in more accurate conclusions as a lack of models increases the likelihood of human error. Because the range of heuristics is so large, the association process between cues and heuristics becomes ambiguous, and ultimately, counterproductive, as the goal of frugal heuristics is to minimize the decision making process. With this in mind, I ask, how do we know when the moment is right and how do we know what connections to look for in developing our algorithms, such as is demonstrated in the decision tree of Figure 1? Does using ambiguous, but simple heuristics result in logical fallacies or otherwise false conclusions?

    Additionally, heuristics are unable to account for real life decisions. Similar to echo, heuristics does not account for emotions or subjective experiences. As it relates to application in the real world, how likely is it that people will put their own lives or the lives of others in the hands of a model that does not account for a plethora of decision making factors? Regardless even if a final decision is theoretically optimal, inevitable biases and personal preferences surpass problem solving algorithms. If I were a doctor why would I trust an algorithm with no room for marginal values, that is those values outside of what cues are present, over my own professional experience and subjective experience of a situation?

Leave a Reply