Simon’s scissors: meta-heuristics for decision-makers

Abstract

Purpose

Are there smart ways to find heuristics? What are the common principles behind heuristics? We propose an integrative definition of heuristics, based on insights that apply to all heuristics, and put forward meta-heuristics for discovering heuristics.

Design/methodology/approach

We employ Herbert Simon’s metaphor that human behavior is shaped by the scissors of the mind and its environment. We present heuristics from different domains and multiple sources, including scholarly literature, practitioner-reports and ancient texts.

Findings

Heuristics are simple, actionable principles for behavior that can take different forms, including that of computational algorithms and qualitative rules-of-thumb, cast into proverbs or folk-wisdom. We introduce heuristics for tasks ranging from management to writing and warfare. We report 13 meta-heuristics for discovering new heuristics and identify four principles behind them and all other heuristics: Those principles concern the (1) plurality, (2) correspondence, (3) connectedness of heuristics and environments and (4) the interdisciplinary nature of the scissors’ blades with respect to research fields and methodology.

Originality/value

We take a fresh look at Simon’s scissors-metaphor and employ it to derive an integrative perspective that includes a study of meta-heuristics.

Citation

Marewski, J.N., Katsikopoulos, K.V. and Guercini, S. (2024), "Simon’s scissors: meta-heuristics for decision-makers", Management Decision, Vol. 62 No. 13, pp. 283-308. https://doi.org/10.1108/MD-06-2023-1073

1. Introduction

“Actions speak louder than words, that is, what people do and get done is more important than what people say, how they talk, or how they present themselves.” This rule-of-thumb comes from a manager. Here is one from another manager for tasks of change-management: “Uncertainty creates fear. Always paint a picture of how it is going to be, what is going to happen.” Finally, consider how a healthcare professional decides whether to go and see a patient when she/he receives a call from an assistant doctor, by using a very simple decision tree, shown in Figure 1.

Those rules-of-thumb are heuristics. The healthcare professional follows an algorithmic procedure, consisting of yes-no questions as input-variables, and a limited number of precisely defined actions as output-variables. The managerial rules are qualitative statements without any specific input-variables, suggesting actions without constraining them to a set.

The fast-and-frugal heuristics research program (Gigerenzer et al., 1999) investigates people’s repertoire of such heuristics, called the adaptive toolbox. What is in the toolbox? This question is key in management and beyond. For instance, leadership is the daily job of doctors and pilots, who oversee assistant-physicians and aircraft-crews, respectively. In business, marketers and strategists develop, structure, and communicate novel ideas, and in academia, researchers do the same. Could there be meta-heuristics, that is heuristics for coming up with new heuristics, regardless of the (e.g. professional) domain and specification (e.g. algorithmic or qualitative) of the heuristics?

This article makes three contributions: To (1) derive meta-heuristics for discovering heuristics, we (2) propose an integrative definition of heuristics (3) using insights that apply to all heuristics. Comments on the scope, style, and goals of this article are presented first.

1.1 Old and new conceptualizations

In formulating meta-heuristics, we enter both known and novel territory. Herbert Simon, the founding father of research on heuristics, studied how heuristics could aid scientific discovery (e.g. Langley et al., 1987), and others following Simon’s footsteps (e.g. Lenat, 1982) investigated meta-heuristics. Our approach stands in that Simonian tradition (see also Katsikopoulos et al., 2024). Yet, for Simon and colleagues, heuristics were to be studied using computer models – a paradigm that also neo-Simonian research on fast-and-frugal heuristics (henceforth: FFH) embraced. But computer models may not be the only tool for conceptualizing heuristics, especially when it comes to heuristics for ill-structured tasks entailing uncertainty, as it abounds in business, science, and other real-world domains (Simon and Newell, 1958). Fitting to such ill-structured, uncertainty-fraught domains, here we also use verbal, qualitative conceptualizations of heuristics.

In so doing, we follow Polya, with whom Newell – Simon’s friend and collaborator – had taken courses at Stanford University (Simon, 1996a). Polya (1945) developed heuristic guiding-principles for mathematical problem-solving. Moreover, he saw ancient origins of heuristics and linked heuristics to proverbs.

1.2 Narrating heuristics

As we elaborate the concept of heuristics, we also conceive of them as narratives. A narrative is a story, or the representation of a story by art. Narratives tell how decisions happened, at least as things appear from the perspective of the storyteller.

Narratives can help gain and represent insights. The act of discovering insights was for Popper (1935/2002, p. 8), one of “creative intuition.” Supporting intuitions about heuristics is the goal of our narratives – including artwork.

1.3 Heuristics in management

In placing this article in a management journal, we aspire to contribute to, integrate and eventually achieve conceptual clarity in research on heuristics in management and other business-contexts (see also Hatchuel, 2023). In those contexts, both qualitative and quantitative heuristics are studied (e.g. Luan et al., 2019; Manimala, 1992). Some research also aspires to undercover the processes by which heuristics emerge (e.g. Atanasiu et al., 2023; Bingham and Eisenhardt, 2011; Guercini and Freeman, 2023).

However, although business-contexts are obvious areas of application for heuristics, and notwithstanding increasing interests in the emergence and conditions of use of heuristics (Artinger et al., 2015; Gigerenzer et al., 2022), much remains to be discovered about the creative processes of discovering managerial heuristics. In fact, such discovery may be hindered by diverging conceptualization of heuristics, with the literature splitting not only along the qualitative-quantitative dimension, but also by how the term ‘heuristic’ is used (see Lejarraga and Pindard-Lejarraga, 2020, for a discussion). Are heuristics simple yet adaptive tools whose rationality is ecological – a (neo)-Simonian view (Gigerenzer et al., 1999; Simon, 1990)? Or do heuristics represent error-prone cognitive shortcuts – a view that is often associated with the heuristics-and-biases program (Kahneman et al., 1982)? While sometimes such radically different conceptualizations are jumbled together (e.g. Hopp and Spearman, 2020, footnote 8; Lopes, 1992), on other occasions they are contrasted – without any theory integration taking place in either case. Literature analyses reflect that rift, too: For instance, Hodgkinson et al. (2023) document increases in publications on “top managers’ heuristics and cognitive biases” (p. 1062; Italics added; capitalized and in bold in original), whereas Loock and Hinnen (2015) review neo-Simonian adaptive heuristics. Diverse usages of the term ‘heuristics’ have also been examined in other business-contexts, for example in marketing (Guercini, 2023a, 2023b; Guercini and Freeman, 2023), international management (Guercini and Milanesi, 2020), customer-supplier relationships (Guercini et al., 2015), and operations management (Katsikopoulos, 2023). Finally, not all heuristics are even called ‘heuristics’ in the first place. How can one navigate such seemingly different views on heuristics? One way forward may consist, precisely, in uncovering meta-heuristics, using common principles underlying all heuristics.

1.4 Overview

In this article, we take a tour d’horizon. We describe heuristics for tasks from management to writing and warfare. We not only discuss heuristics from the scientific literature, but also describe heuristics in religious texts, folk-wisdom, proverbs, fairy-tales and in practitioner-accounts, including our personal experiences as research-practitioners.

In Section 2, we propose an integrative definition of heuristics. In Section 3, we present 13 meta-heuristics and explain how they relate to 4 common principles, grounded in the notion of bounded rationality, put forward by Simon (e.g. 1955, 1956) and developed further into the notion of ecological rationality (Goldstein and Gigerenzer, 2002) within later neo-Simonian research on FFH. We close with a brief discussion in Section 4.

2. Towards a definition of heuristics

2.1 Neo-Simonian heuristics

The notions of bounded and ecological rationality are captured by a metaphor, coined by Simon (1990, p. 7): “Human rational behavior … is shaped by a scissors whose two blades are the structure of task-environments and the computational capabilities of the actor.”

Simon’s scissors-metaphor reflects key insights on heuristics. Behavior and performance emerge from the interplay between minds and environments: they are not just a function of minds, but of how those minds nestle into context. Moreover, bounds in information-processing capacity, knowledge, and time may not imply mediocre performance. Rational behavior can emerge if cognitive mechanisms match the structure of task-environments.

Those and other insights captured by the scissors-metaphor fuel our development of meta-heuristics. We posit that those insights also underpin neo-Simonian research on FFH, which is the theoretical framework in which we embed our propositions. FFH are simple, ecologically grounded models of bounded rationality that have been applied to diverse task-domains. A principal empirical finding is that heuristics can lead not only to quick (hence: fast) decisions based on minimal information (hence: frugal), but also to smart ones: On several criteria, ranging from accuracy to transparency, heuristics can outperform more complex, information-greedy strategies (e.g. Gigerenzer and Gaissmaier, 2011). The decision tree shown in the introduction is an example of a class of FFH dubbed fast-and-frugal trees (Katsikopoulos et al., 2020).

In what follows, we present and eventually update the FFH-framework’s definition of heuristics. The update addresses an implicit broadening of the FFH-concept over the past decade.

2.2 What are heuristics I? Implicit broadening of the notion of FFH

In the spirit of Simon’s earlier work, in the nineties the FFH-program started out by studying heuristics as computer models of information-processing. Those were characterized by algorithmic rules prescribing, for example, the order in which information is searched for and when search stops. Gigerenzer and Gaissmaier (2011) define: “A heuristic is a strategy that ignores part of the information, with the goal of making decisions more quickly, frugally, and/or accurately than more complex methods” (p. 454).

Examples of complex methods of information-processing are statistical and machine learning tools such as regression, Bayesian networks, or random forests. Those perform many more computations than heuristics. To illustrate, a regression can weight and integrate several variables, but the take-the-best-heuristic (Gigerenzer and Goldstein, 1996) orders variables and considers them sequentially, deciding based on just one; and at the other extreme of simplicity, the tallying-heuristic (e.g. Dawes, 1979) counts (sums) all variables but does not weight them.

Ordering and counting are traditional examples of information-processing building-blocks of heuristics (Gigerenzer et al., 1999). However, humans are capable of other operations, too – such as listening, speaking, or moving. Increasingly, FFH relying on such building-blocks are discussed. Examples include heuristics for managing employees, such as “First listen, then speak” or “If a person is not honest and trustworthy, the rest doesn’t matter” (Gigerenzer, 2014, p. 117). These heuristics differ from those that have been studied in the first decades of research on FFH (e.g. Gigerenzer and Goldstein, 1996; Schooler and Hertwig, 2005): they are not computational models of information-processing. Rather, they are simple, qualitative principles. In this, however, they do resemble numerous other heuristics; notably rules formulated by practitioners in different areas, including ourselves as research-practitioners. Here we introduce, for instance, one of our own guiding-principles for writing, dubbed take-the-last, in addition to discussing guiding-principles we have found useful for coming up with novel heuristics, our own meta-heuristics.

Indeed, such simple, qualitative guiding-principles are often of somewhat idiosyncratic nature: they are developed by practitioners – be they scientists or managers – as those practitioners tackle problems in their respective fields of expertise. That is, those heuristics typically emerge from personal experiences, and different practitioners, working in the same domain, might come up with different heuristics. Those experiences might, moreover, be shaped by a practitioner’s unique features, including memories of other experiences, stories told to her/him, or her/his character. Also in that regard those qualitative guiding-principle can differ from earlier computational models of heuristics which were often treated as more general models, hypothesized to describe the behavior of larger (or all) groups of people (see e.g. Bröder’s, 2000, discussion of take-the-best as “universal hypothesis”, p. 1332). One can roughly relate those two types of heuristics – idiosyncratic versus general ones – to the distinction between idiographic and nomothetic approaches to knowledge (Hurlburt and Knapp, 2006, p. 287).

Importantly, despite their qualitative fuzziness and possibly idiosyncratic nature, those actionable principles can nevertheless reflect insights about the world. By insights we mean (1) knowledge about the external environment that could generalize beyond an individual’s unique experiential environment, as well as (2) knowledge about one’s internal self, which could generalize beyond one’s own momentary specific condition.

The former type of generalizable knowledge may tap onto what has been called local evidence in evidence-based management-education: “causally interpretable data that is collected on-site in companies to address a specific business problem” (Dietz et al., 2014, p. 397). Such local insights may generalize both within the same specific business-context as well as across others, be they in business or elsewhere. The latter type of knowledge can generalize either in time but still refer to the same individual (e.g. its future self) or to other individuals.

Such generalizable knowledge can be produced by research but does not need to; insights may come from personal experiences and subsequent introspection. For instance, first-listen-then-speak and other leadership-heuristics were gathered by interviewing top-managers (Gigerenzer, 2014); and take-the-last was gleaned from our writing experiences as scientists. These heuristics communicate and make actionable for others a single practitioners’ insights. Once articulated, such personal insights can then also be subjected to systematic empirical research.

2.3 What are heuristics II? Heuristics in management and other fields

How can the notion of heuristics as computational – and possibly ‘nomothetic’ – models be integrated with one that is not only qualitative, but that can also be highly idiosyncratic? And could there be room for combinations of those two and/or other conceptions?

Consider how Simon (1990, p. 2) summarizes Pasteur’s germ theory of disease: “If you observe pathology, look for a microorganism-it might be causing the symptoms.” This qualitative statement expresses a qualitative guiding-principle for action (look for a microorganism) that reflects an insight about the world (microorganisms might cause pathological symptoms) and that can easily be put into action to find a good solution to a problem (what is the reason for observing this pathology?). It is a qualitatively cast heuristic. Yet, this heuristic is not idiosyncratic.

Or re-consider the healthcare professional’s procedure (Figure 1). This fast-and-frugal tree is, likely, rooted in an idiosyncratic personal professional context; yet, contrary to the two managers’ qualitative rules we introduced in the very same paragraph in the first lines of this article, fast-and-frugal trees are algorithmic models, readily studied in computer simulations (Katsikopoulos et al., 2020).

Now let us turn to qualitative experienced-based heuristics. Such heuristics abound in business-contexts. For instance, Bingham and Eisenhardt (2011) examined how heuristics learned through experiences in other countries are used by firms entering new foreign markets. Those firm-level heuristics correspond to qualitative guiding-principles, some of which seem applicable to specific industries (“Enter countries with lots of pharma activity”, p. 1444) and others potentially generalize across industries and firms (“Work with experienced local country manager”, p. 1445). Or consider Sutton’s (2007) advice on how to recognize toxic employees. (“Does the alleged asshole aim his or her venom at people who are less powerful rather than at those people who are more powerful?”, p. 8). As another example, Tuckett and Nikolic (2017) report heuristics from fund-managers; one is “Look for shares hit by possibly exaggerated rumors (e.g. of impending litigation or compensation pay-outs)” (p. 507).

Qualitative experienced-based heuristics can also be found in management-research that explicitly grounds itself in the FFH-framework. By conducting in-depth interviews and ethnographic observations with the management of textile-manufacturers, Guercini (2019) observed heuristic rules in marketers' narratives, including “multipliers” (p. 196) as in mark-up-based pricing, “thresholds” (p. 196) as in goal setting, or “calends” (p. 197) as in choosing the right time to decide. Interviewing investment and central bankers, Ehrig et al. (2021, p. 1753) discovered qualitative heuristics for strategic interaction in financial markets (e.g. “Communicate a coherent narrative”; “If everyone agrees, be skeptical”).

Finally, consider shapeshifting heuristics: Gigerenzer et al. (2022) describe the very idiosyncratic heuristics used by Musk and Bezos, respectively, for hiring decisions. One question Bezos asked is whether he would “admire this person” (Gigerenzer et al., 2022, p. 176). Yet most personal heuristics, though linked to their inventors though the experiences made, are not limited in their potential usefulness to them and also Bezos’ question might serve other employers analogously. What is more, Gigerenzer et al. show how Bezos’ and Musk’s heuristics can be expanded into quite general fast-and-frugal trees, and such transition from qualitative personal experience into quantitative, general heuristics has also been documented for other areas (e.g. medicine: Gigerenzer, 2007, pp. 169–178).

This, however, is not to say that there might not be heuristics that are both derived from personal experience and only useful – in the extreme – to their inventor. For example, personal characteristics – character, skills, worldviews – might permit a manager to act in ways others cannot imitate. Here we touch upon insights about oneself: Know thyself! An integrative definition should also offer room for such extremely idiosyncratic heuristics.

2.4 What are heuristics III? An integrative view on heuristics

To allow for integrating diverse types of heuristics into one framework, we propose the definition of Table 1. That definition refers to all behavior, rather than only decision-making as in Gigerenzer and Gaissmaier (2011). Further, note that the definition aligns with Lenat’s (1982). According to him, all heuristics reflect insights; heuristics are “useful for guiding” (p. 192), “a piece of knowledge capable of suggesting plausible actions to follow ….” (p. 192). He stressed that “for a body of heuristics to effective … each heuristic must specify a … context in which its actions are especially appropriate …” (p. 192).

Table 1’s definition integrates computational and qualitative conceptualizations. It also allows for highly idiosyncratic experiences to shape heuristics – be it actionable knowledge that could generalize to be useful for others or knowledge that might be useful only to a single individual.

The definition addresses a growing gap between the practice of research on FFH and its theoretical foundations. Particularly, in the management-literature, increasingly heuristics are dubbed ‘fast-and-frugal,’ without the studied heuristics corresponding to the computational models developed originally within the FFH-framework (see Gigerenzer et al.’s, 2022, p. 192, call to “translate verbal rules into algorithmic models”). For instance, De Treville et al. (2023) describe how Toyota’s practices, such as jidoka, andon and kanba, can be conceived of as FFH for production (exploitation), whereas gemba, kaizen and five-whys can be thought of exploration-heuristics. Whereas some Toyota practices might be cast into algorithms – we provide an example below – it is difficult to see how this could be done for all. Without an updated definition, the label fast-and-frugal risks becoming arbitrary, applicable to any procedure that qualifies, by some standard, as simple, fast or frugal. At the same time, an updated definition may lead to further questions, such as which types of heuristics can be algorithmically cast, and which ones not – and why. For example, the fuzzy nature of qualitative, verbal heuristics could correspond to the fuzzy nature of the ill-structured problems to which they apply, and this “lack of definiteness could be functional: One may speculate that fuzziness allows for those narrated descriptions of tools and environments to attach to different kinds of ill-structured real-world situations a decision-maker may experience” (Marewski, 2023, p. 55).

Importantly, Table 1’s definition explicitly distinguishes heuristics from notions that, though embodying some form of simplicity, speed, or frugality, cannot explain smart behavior. Fundamental to research on FFH is to uncover under what conditions a heuristic enables smart behavior. Hence, to qualify as heuristic, a notion must come with a hypothesis about environmental fit. In the heuristics-and-biases literature, such fit is, commonly, not part of what is dubbed a heuristic. Rather heuristics are conceived of as the mind’s product, leading to biases and other irrational behavior, with the yardsticks for rationality not being ecological, but context-general norms (e.g. rules of logic, utility-maximization; Gigerenzer, 1996). In the past, FFH and heuristics-and-biases could be set apart also by consistent differences in their specification, with heuristics-and-biases typically being qualitatively cast. Yet, qualitatively cast heuristics-and-biases must not be put in the same basket as the emerging body of qualitatively cast of adaptive heuristics. Only if the former come with hypotheses about environmental fit they qualify as heuristics following Table 1’s definition.

Conversely, this definition aids identifying heuristics also when those describing a way of acting do not think of it as heuristic and/or call it differently. Indeed, in the literature the term heuristic is not consistently used. For instance, Eisenhardt, and coworkers rather refer to simple rules instead (Sull and Eisenhard, 2015). And a practitioner, interviewed about how she/he makes decisions, may not even know the term ‘heuristic,’ and if it is just that this term is uncommon in her/his professional culture (few French-speaking managers use the term ‘heuristique’). Finally, simple guiding-principles that reflect insights about the world, as they can be found in folk-wisdom, will almost never be called ‘heuristics’; rather it is upon the researcher to decide whether a statement qualifies as heuristic.

Indeed, we believe that numerous heuristics can be found in proverbs (“Do not put all your eggs in the same basket”, Hafenbrädl et al. 2016, p. 218), ancient books, including religious texts (“Do to others as you would have them do to you”, Luke, 6:31), as well as in other narratives. And that across cultures and time. For example, a lemma the Sioux (Lakota) transmitted to their youngsters was “It is better to die on the battlefield than to live to be old” (Hassrick, 1992, p. 47, translated; for a first-hand account, see Standing Bear, 1975, p. 124, p. 135). An example of a battle-tactic, attributed (e.g. by the later military writer, Vegetius, p. 107) to Scipio Africanus, the Roman general who beat Hannibal and put an end to the second Punic War, is: “An escape-route should be offered to the enemy so that they may be more easily destroyed in full flight.” Or consider fairy-tales, such as Hänsel and Gretel, or The Boy-Who-Cried-Wolf: Those stories can convey rules-of-thumb to children (Marewski and Hoffrage, 2020): “Do not trust strangers!” (p. 304) and If you lie once, nobody will believe you when you speak the truth!, respectively. Throughout history, a function of such narratives may have been to transmit actionable insights for being effective in the world. The world is characterized by uncertainties. We cannot know for sure what others’ intentions are, whether we will be healthy, what works on the battlefield, but a repertoire of simple guiding-principles may aid. Heuristics are tools for shaping behavior – that is, adapting to, and surviving, even thriving – in a world in which not everything is calculable.

Our proposed definition permits discovering heuristics through narratives; this way addressing a call by Gigerenzer (2023) to discover how narratives and heuristics relate to each other. For instance, he wonders if “Heuristics select narratives to explain events, and ‘big’ narratives select the heuristics that people live by, to execute their values and moral principles” (p. 43). We believe narratives may aid decision-makers to understand uncertain environments, particularly to recognize patterns in ill-structured problems, which may help them to select their decision-strategies (Marewski, 2023). Narratives may also serve to explicate heuristics from one practitioner to another and also we as research-practitioners use stories to explicate our respective idiosyncratic heuristics to others, including to the readers of this article. Finally, the telling of narratives itself may lead to new heuristics. For instance, to identify heuristics adopted in specific business-contexts, “tales from the field” can be used (Guercini, 2019, p. 119; see also Guercini 2023b), and our writing of this paper shaped – through writing-heuristics (e.g. take-the-last) – our ideas on meta-heuristics.

That said, note that, in the definition, not every simple principle, even if it comes with a narrative on its environmental fit, qualifies as heuristic. If the hypotheses about insights a candidate heuristic operates on do not pass empirical tests, then one cannot speak of insights about the world, and by extension, not label a simple statement a heuristic.

2.5 What are heuristics IV? An integrative view on meta-heuristics

We are now just a step away from defining meta-heuristics. They are also heuristics (Lenat, 1982). Just like all other heuristics, they can be quantitative or qualitative guiding-principles. They may, too, have either a more idiosyncratic or a more generalist scope (Talbi, 2009). And regardless of their nature, they can be transmitted through narratives (e.g. personal stories, scientific articles). The only difference between heuristics and meta-heuristics is that the latter apply to the task of discovering heuristics, and thus reflect insights about heuristics and ways to discover them, as opposed to insights about other domains of expertise (e.g. management, warfare). Thus, we posit the definition in Table 1.

Both definitions, that of heuristics and that of meta-heuristics, emphasize the distinction between uncertainty and risk (Knight, 1921). What does that mean? Under risk, all alternatives, their consequences, and the probabilities of these consequences are known. In such environments, given sufficient computational power, it might be possible to optimize – identify the utility-maximizing action. In contrast, uncertainty refers to situations “where the exhaustive and mutually exclusive set of future states … and consequences …” (Gigerenzer, 2021, p. 3548) are not known or knowable – instead, there can be surprises, and thus optimality is out of reach. Uncertainty requires heuristics. How to find them?

3. Towards a formulation of meta-heuristics to discover new heuristics

For a research-practitioner starting to write an article, staring at that discouragingly empty document, can be overwhelming. What arguments to make? How to structure them? Writing comes with myriads of possibilities; the practical challenge is to decide for a good one. Here is a heuristic one of us adopts: Take-the-last: Start a new sentence with the word or notion on which the preceding sentence ended.

This heuristic does not allow designing text upfront; it does not require reflecting on structure or to otherwise plan; it just leads to a structure and ultimately a text. How exactly does take-the-last work? Figure 2 illustrates Simon’s scissor metaphor by narrating how one of us experiences writing with take-the-last.

Simon’s scissors metaphor serves, in what follows, to articulate general insights on heuristics. All heuristics – including our meta-heuristics – are grounded in those insights, which we subsume under four principles. They concern the (1) plurality, (2) correspondence (3) and connectedness of heuristics and environments and (4) the interdisciplinary nature of the blades with respect to both fields of research and methodology.

3.1 Principle of plurality

From Scipio’s Offer-escape-routes!, the Christian’s Treat-others-like-oneself!, and the Boy-Who-Cried-Wolf’s Do-not-lie! to take-the-last: Heuristics are incredibly diverse. They come with their own narratives, explicating the heuristics further, be it stories of the battles of Roman legions, led by the Republic’s war hero, Scipio; a narrative of Jesus giving a speech to his disciples; a fairy-tale of a boy, messing with the inhabitants of his village; or scientists, trying to share with each other (and their students) their respective tricks for writing. Many of them are qualitative, like those just listed, but as the healthcare professional’s tree (Figure 1), others come as algorithmic models. Some heuristics read like general principles; others come across as more specific. Certain guiding-principles take a more metaphorical disguise; others such as Simon’s summary of Pasteur’s germ theory of disease are concrete if-then statements.

Are environments of heuristics similarly diverse? Early research on FFH focused on well-structured inference-tasks, such as judging which of two cities has more inhabitants (Gigerenzer and Goldstein, 1996). Such tasks come with a matrix of objects, their attributes, and criterion values –a quantitative environment (Katsikopoulos, 2011; Martignon and Hoffrage, 1999). Take-the-best and tallying are computationally-specified heuristics for such environments, with regressions and other statistical tools being direct competitors.

Let us now consider some quite different couples of blades, zooming into the heuristic first-listen-then-speak. As Gigerenzer (2014) points out, first-listen-then-speak can support leadership-tasks in business and cockpits. But it can also support leadership in other environments, ranging from PhD supervision to medicine. The idea is that when a senior, more powerful leader (e.g. a CEO, pilot, professor or chief-physician), works together with a more junior colleague (a lower-level manger, younger co-pilot, PhD student, assistant-physician), by letting the junior speak up first about what action to take (e.g. what treatment to propose for a patient, how to react to an incident on a flight), the senior avoids influencing, and potentially shutting up, the junior. Letting the junior speak first provides a check and potentially a learning opportunity to the senior; it also guards the senior against embarrassment or insistence on mistaken ideas. Here, the environmental blade is comprised of people, standing in hierarchical relations to each other, facing a task. The environment may come with diverse costs to the senior (e.g. losing face) and potentially also to the junior and others (e.g. passengers of an aircraft who might die in case of a mistake).

Such ill-structured social environments differ from the quantitatively-specified ones discussed previously. The social environments are narrated. Notably, qualitatively-cast heuristics gain meaning from descriptions of their environments, and often such accounts of environments and their heuristics may come as stories, told by friends, colleagues, the news, and so on, prior to that a heuristic and a more general description of its environment may be constructed from the story. In short, the space of couples of blades is rich. Plurality reigns.

Insight – principle of plurality

Environments and heuristics are diverse; there is more than one couple of blades, and these can be quite different from each other.

Based on the principle of plurality, one can come up with various meta-heuristics; Table 2 provides two examples. Neo-Simonian FFH-research has followed meta-heuristic-1: Gigerenzer et al. (1999) proposed that people rely not on a general, one-size-fits-all-method for decision-making but instead employ a toolbox as a function of the environment. Contrasting this adaptive-toolbox-approach, others put forward the thesis that behavior is best modeled by an all-purpose decision-making mechanism (expected-utility-maximization, parallel-constraint-satisfaction; Glöckner and Betsch, 2008). Meta-heuristic-1 has also fueled skepticism about that latter thesis, scientific debates, and adversarial collaborations (Marewski et al., 2018). Meta-heuristic-2, in turn, is consistent with experimental and modeling work striving to understand how people select among different heuristics as a function of environments (Rieskamp and Otto, 2006).

3.2 Principle of correspondence

The following is a narrative of a (re)discovery (see also Katsikopoulos et al., 2024) that started with a critical remark: An eminent scientist, visiting the ABC-Research-Group when Author-1 started his PhD studies, pointed out that the ABC-Research-Group was lacking a theory of environments! Back then, Author-1 believed that this scientist was right.

But many years later, looking at the reflection of a mountainous landscape on the surface of Lake Geneva, a thought struck Author-1: According to the principle of plurality, each heuristic must be its own guide for discovering the way to develop a theory of its environment!

“[S]erving to find out or discover” (Gigerenzer and Gaissmaier, 2011, p. 454) is what the Greek word heuristic means: Following the Roman engineer and author of De architectura, Vitruvius, “εύρηκα … I have found it out” (Book IX p. 205), are the words. Archimedes had cried out (many years earlier), namely after he, while taking a bath, discovered another principle of correspondence: regardless of how irregularly shaped it may be, the volume of any item corresponds to the volume of liquid – water – it displaces. Archimedes had tried to solve a specific problem: understanding if the King’s crown was made of pure gold (Vitruvius, Book IX pp. 204–205). Archimedes’ principle follows displaced water to find the volume of objects. Cognitive scientists can follow any mental blade and find environmental blades – and vice versa.

Insight – principle of correspondence

The defining characteristics of an environment (e.g. its elements) are, in one way or the other, inherent in heuristics (e.g. what input it operates on, what assumptions it makes about that input). Correspondingly, theories about environments can be informed by theories of heuristics, and vice versa.

But what does ‘inherent in one way or the other’ mean? The principle of correspondence embodies the insight that there is no single mapping between heuristics and environments. Rather the nature of the mapping varies and depends on the specific heuristic-environment-couple. For instance, tallying counts. Since tallying prescribes precise computations, the mapping will follow that logic, and a theory of tallying’s environment will include, for example, quantities of items that an agent would be able to separate from each other and can be enumerated. In contrast, first-listen-then-speak may lead researchers to a theory of environment that will feature, among others, agents who can listen and agents who can speak.

The diversity of mappings is also consistent with Table 1’s definition. Recall, heuristics reflect insights about the world. Those insights are hypotheses about how to reach good solutions to problems. As such, they must build on lawful relations in the world – Simon (1990) calls those laws invariants. Invariants are vastly diverse. For one, invariants can be rather precise, but many are approximate, even in the natural sciences. “For instance, the isotopes of the elements have atomic weights that are nearly integral multiples of the weight of hydrogen” (pp. 1-2). Second, while invariants can take a quantitative form – think of Newton’s laws – they can also come as qualitative statements. Simon uses Pasteur’s germ theory of disease as example for the latter. Third, many invariants are specific and not general. This is the case in biology: “Because of inter-species molecular differences, even the important general laws (e.g. the laws of photosynthesis) vary in detail from one species to another (and sometimes among different individuals in a single species)” (p. 2). Commenting on the human species, he stresses: “Many of the invariants we see in behavior are social invariants. And since they are social invariants, many are invariant only over a particular society or a particular era, or even over a particular social or professional group within a society” (p. 16). A final dimension comes with the nature of change over time in a system. Change can be simple and “invariant laws” (p. 2) be captured by differential equations – Simon uses the “movements of the heavens” (p. 2) as example – or it can be adaptive. Adaptive change, as it can be found in human and other biological systems, and “...which is as much governed by a system’s environment as by its internal constitution, ... [makes it] difficult to identify true invariants” (p. 2) - truly ‘invariant’ laws. In short, because invariants of the world are diverse, insights about those invariants are diverse, and as consequences heuristics, their environments, and the mappings between each of the blades must be diverse, too. The principle of plurality, introduced above, comes with plurality in correspondence!

How much plurality is there? We believe that Simon’s (1990) dimensions of invariants offer a coordinate system for systematically describing the adaptive toolbox (see Gigerenzer and Gaissmaier, 2011, for a call to develop corresponding integrative theory). Consider folk-wisdom about human behavior. A person who lies once will likely lie more often reflects a qualitative, approximate invariant; but it may be rather specific, likely depending on culture or other variables, such as personality or the group-context. Similar coordinates may apply to first-listen-then-speak, which reflects an invariant, likely equally malleable by culture, that humans tend, against better knowledge, to follow actions suggested by authorities. Now consider the invariant reflected in Scipio’s heuristic to build escape-ways for one’s enemies. This invariant is described in Vegetius’ Epitome of military science: “The only hope of safety for the defeated is to expect no safety” (p. 108), because if there is no way to escape, the only chance to survive is to fight. In contrast to the former two, this invariant is precise, and it is also general, because the behavior to fight if trapped can also be observed in animals. Quantitative laws, in turn, may underlie yet other heuristics, and so on.

Returning to our narrative, with the principle of correspondence, Author-1 ‘re-discovered’ Simon’s insights about invariants and learned to understand the scissors-metaphor in an – to him – unexpected, novel way: The principle of correspondence does not represent a theory of environments, but it allows formulating meta-heuristics for discovering the environments of any heuristic.

Indeed, the FFH-program has followed, successfully, meta-heuristics-3 and 4 (Table 2) and has been able to benefit from and contribute to various disciplines (Katsikopoulos et al., 2020). Meta-heuristic-4 reflects the ecological rationality dimension (Todd et al., 2012).

It is also those two-meta-heuristics that set the FFH-framework apart from Kahneman et al.’s (1982) heuristics-and-biases-program. Underspecified notions such as availability and representativeness can jumpstart research. However, a lasting lack of precision can hinder progress (Gigerenzer et al., 1999), not only when it comes to describing the mental blade, but also when it comes to uncovering its environmental blade – which is missing in the heuristics-and-biases-program. Ironically, here the principle of correspondence applies, too. Looking at a researcher’s theory about the mind, affords making inferences about that researcher’s theory of environments.

3.3 Principle of connectedness

One can take the coupling of heuristics and environments further and suggest that heuristics and environments should be understood as systems. Plurality and correspondence come with connectedness! That insight can be introduced by three observations.

First, in his The Sciences of the Artificial, Simon wrote “Human beings … are quite simple. The apparent complexity of our behavior over time is largely a reflection of the complexity of the environment in which we find ourselves” (1996b, p. 53). Simon considered an ant moving on a beach: when he drew the ant’s path on paper, it looks “irregular, complex, hard to describe” (p. 51). But this complexity is “not a complexity in the ant,” (p. 51) he said, but one inherent in the “wind- and wave-molded beach” (p. 51). Figure 3 illustrates Simon’s insight of how the environment shapes the complexity of an ant’s behavior.

Second, in this example, the ant’s behavior depends on the environment. But an organism can also shape its environment, with that environment, conversely, acting back onto the organism, leading to dynamic system with emergent properties. Ant hills, piled up over time, are such a case: A past action of one ant, placing a fir needle at one place, conditions future actions and the environment of that and other ants. Or consider how take-the-last creates text-environments: each sentence conditions the next sentence’s contents. This way chains of paragraphs emerge (Figure 2).

Third, the contents of those sentences are not just conditioned by what is already written (i.e. the last word); contents also depend on what concepts are in memory. Said Simon (1996b): “Long-term memory operates like a second environment, parallel to the environment sensed through eyes and ears, through which the problem solver can search and to whose contents he can respond” (p. 88). Something similar holds true for ants: their behavior is not only shaped by physical surfaces but by percepts of the environment. For example, Buffon’s needle, a heuristic operating on pheromone marks laid by ants, likely informs ants’ choice of where to construct a new nest (Mugford et al., 2001). How the world is perceived and remembered matters for what we do: heuristics operate on input from the perceptual system and memory, constituting internal representations of external environments, or to use Simon’s (1996b, p. 25) term: “inner environment[s].”

Insight – principle of mind-environment connectedness

Heuristics and environments do not work in isolation or parallel, but in a mind-environment system. That system lets cognition (e.g. precepts, thoughts), behavior (e.g. moving part of one’s body) and performance emerge.

How can one study such mind-environment systems? There are two complementary meta-heuristics, we dub micro (No. 5) and macro (No. 6; Table 2). What might Simon have had to say about micro? He saw both humans and machines as information processors and believed that their behavior could be understood similarly. Together with Newell, Simon paved the ground for modern-day AI and developed information-processing computer models (e.g. the logic theory machine in the fifties). As Simon (1990, p. 13) put it: “The whole congeries of mechanisms of human rationality must somehow be organized in the human brain to work together in a coordinated fashion.” In that article, Simon points to one class of computer models, closely associated with Newell’s name, as means to study mind-environment systems: Unified theories of cognition, also called cognitive architectures, that integrate theories of memory, perception, motor-action and other components of cognition into a single large-scale computational model. Through simulation, one can then trace how the different components of the modelled cognitive system interact in time and space, and eventually better understand each such parts, and the dynamic, emergent properties of their joint interplay.

For example, ACT-R (Anderson et al., 2004), an architecture that comes with a strong environmental grounding, has been used to model that interplay: A person’s ability to rely on (1) the recognition and (2) fluency heuristic, respectively, hinges on (1) the probability and (2) speed of retrieving a memory record of an object, which in turn, are shaped by patterns of encounters with the object. Simulations showed how forgetting boosts the accuracy of those heuristics (Schooler and Hertwig, 2005). ACT-R allowed, furthermore, uncovering how cognitive niches channel when people can execute the fluency, recognition and other heuristics (Marewski and Schooler, 2011). Those niches are sets of affordances that emerge from how external and inner environments interplay – such when patterns of environmental occurrences of objects give rise to encounters and recognition memories, with memory retrieval affording a person to rely on recognition in decision-making.

Let us now turn to macro. As Simon (1996b) put it: “The outer environment is defined by the behavior of other individuals, firms, markets, or economies” (p. 25). For example, Berg et al. (2010) used agent-based simulations to model how a heuristic that operates on the contents of recognition memory may shape, at a macro-level, neighborhood segregation. Meta-heuristic-6 is consistent with such work.

Furthermore, both micro and macro-level thinking can motivate research into the performance of heuristics: Consistent with meta-heuristic-6, Spiliopoulos and Hertwig (2020) investigated the ecological rationality of using heuristics in strategic interactions. In line with meta-heuristic-5 and 6, Stevens et al. (2018) ran agent-based simulations to study how social contact patterns among agents may buffer costs of forgetting in the evolution of cooperation. Those agents used different strategies, including heuristics such as tit-for-tat.

Finally, consider scientific discovery. We believe many meta-heuristics do not work in isolation, but as a system. For instance, meta-heuristic-1, which prescribes skepticism towards all-purpose decision-making mechanisms and calls to develop theories of heuristic selection, works well with meta-heuristic-13, prescribing to collaborate with researchers’ who disagree with oneself. That latter heuristic, in turn, warrants being exposed to critical thinkers in one’s research-environment. For scientific discovery, one could therefore formulate meta-heuristic-7 (Table 2).

3.4 Principle of cutting-through

Contrary to some stances (e.g. Glöckner and Betsch, 2008), assuming multiple heuristics and environments does not imply theoretical segregation and arbitrariness. What common elements integrate the adaptive toolbox and its environments? Let us turn to the mental blades, and then discuss what integrates the environmental blades.

The FFH-framework suggests that a multitude of heuristics might be composed from a smaller set of building-blocks, analogously to how the entries in chemistry’s periodic table can be decomposed into smaller elements (Gigerenzer and Gaissmaier, 2011). Above we introduced ordering and counting as examples – and we could have added multiplying, too, which underlies many business-decisions (e.g. setting prices through mark-ups; Guercini, 2019). Another is the capacity for recognition, which is key to the recognition and fluency heuristic, respectively. Recognition features, too, in the rule-of-thumb taught to children Don’t-trust-strangers! It also plays a role in another heuristic one of us uses for writing, namely as seeding for take-the-last: Familiar-story-first: Start a paper by writing down an anecdote, an example, a story you are familiar with. And as Simon (1990) puts it, recognition is key to managing “such diverse tasks as grandmaster chess playing, medical diagnosis, and reading” (p. 9). To experience the power of recognition, see the vignette and drawing in Figure 4.

Or consider the building-block of betting-on-the-last: While we have introduced this building-block with our writing heuristic take-the-last (Figure 2), it also forms part of a different heuristic with the same name: When inferring which of two objects scores a larger value on a criterion, Gigerenzer and Goldstein’s (1996) take-the-last identifies the predictor used for the previous inference, and then relies on that predictor to make the current inference. Furthermore, the building-block can be found in forecasting in healthcare and marketing, in the guise of two recency heuristics: “Predict that this week’s proportion of flu-related doctor visits equals the proportion from the most recent week” (Katsikopoulos et al., 2022, p. 614), and “[Predict that] [t]he past 10% ... best customers in a customer base will also be the future 10% ... best customers” (Wübben and Wangenheim, 2008, p. 89). Yet another instantiation of the block emerges in common wisdom: A heuristic is to start searching for a lost item (e.g. a key) at the location of the place where that item was last seen. Finally, consider the Toyota production-system and its five-whys. This method can be thought of as a heuristic for quality management (Figure 5). It helps identify the causes of production glitches by iteratively asking a why-question on the last identified cause, eventually seeking-out the root-cause. In short, surprisingly different heuristics for writing, inference, marketing, healthcare-management, search and quality-management can all be built from one building-block.

Reversely, when seen through the lens of a given building-block, seemingly different environments have common elements relevant for the workings of that block. For instance, betting-on-the-last applies to environments where encountering the most recent element predicts encountering the next one. Recognition-based strategies, in turn, apply to environments where the following two conditions hold: (1) Some items encountered in the past will be re-encountered in some future. (2) Such re-encounters carry informational value (e.g. for making inferences). Statistical structure corresponding to (1) has been described in human and Chimpanzee environments (Stevens et al., 2016), and (2) has been reported for numerous task-domains featuring items from banks to buildings (Goldstein and Gigerenzer, 2002).

Or consider the following leadership-heuristic and its narratives: Commit potential collaborators (e.g. colleagues, allies) to a course of action by closing their escape routes! This heuristic reflects – just as Scipio’s heuristic to build escape-ways (golden bridges) for ones’ enemies – the invariant that, if trapped, humans and animals can become aggressive fighters. The heuristic applies to situations where escape routes can effectively be identified and closed. For instance, ancient Roman and Greek battlelines were organized such that soldiers fighting in the first line could not easily turn around and flee: the subsequent lines physically blocked fighters. In later armies, soldiers waiting behind the lines were tasked to hunt down individuals trying to flee. Cortés, conquistador of the Aztec empire, is said to have deprived his men of escape-routes – by destroying the ships upon his arrival in Mexico. Seals to escape routes do not need to be that obvious: An example is Caesar’s crossing of the Rubicon River. This action implied that Caesar and his men would have no choice but to fight civil war – so, for instance, the narrative of Caesar’s life told by the Roman writer, Suetonius (Divus Iulius, 31) suggests. Indeed, Cortés’ conquistadores allegedly saw the parallel of their leader’s behavior to that of Caesar: there were only two outcomes possible, be victorious or perish (Reynolds, 1959) – historical narratives intertwine. Finally, take the development of one of the most successful modern airliners, the Boeing 777. As a former Boeing employee, Pandey, describes, at a stage in that process, certain (Boeing) “. . . chiefs were reluctant . . . ” (Pandey, p. 83): possibly they feared the responsibility and, in case of failure, the potential blame “for holding up the Early ETOPS approval” (Pandey, 2010, p. 84) – something that could be detrimental to their careers. In one of the meetings with those chiefs, Alan Mulally, “a charismatic leader” (Pandey, 2010, p. 83) who was then in charge, allegedly “. . . did not ask any of his chiefs if they had any questions. Instead, he turned to his chiefs and said, ‘If any of you still have any doubt …, let me know. I will help you find another assignment’. Suddenly the room was very quiet” (Pandey, 2010, p. 84). Did this statement make the 777’s success the chiefs’ only career option? If so, ancient battlelines, conquests and modern-day business-contexts would share an environmental element: identifiable and closable escape routes.

Discovering the common elements of environments and the common building-blocks of heuristics warrants multi-disciplinary expertise, as well as often also practical experience. Learning is broad; it goes in all directions. For instance, for a practitioner it may be easy to see how also modern-day business-to-business environments can exhibit identifiable and closable escape routes, namely when a supplier depends on a major or sole customer. Hence, what we couched a heuristic for leadership above, may also qualify as a heuristic for managing B2B-relations. To give a few more examples, HR-practitioners will easily see how Scipio’s heuristic to build escape-routes applies to their job. Rather than forcing an opposing party to fight a work-place conflict to the bitter end, it can be wise to offer that party ‘ways out’ (e.g. in case of superior-subordinate conflict, offering one party to be transferred to another department). Or consider a psychologist, charged to conduct a relational audit, due to a situation of harassment in an organization: a useful heuristic that reflects the same insights as Scipio’s may be: If there is an implicit threat that at the end of the investigation, one party may loose its job, the truth will be lost first. Finally, consider how the heuristic “If a person is not honest and trustworthy, the rest doesn’t matter,” cuts beyond the business-environment for which Gigerenzer (2014, p. 117) described it: The simple rule to make trustworthiness indispensable matters greatly for a senior physician directing a unit in a large hospital, as he remarked to Author-1. Physicians working in shifts need to entrust patients to each other; thus, if a physician hides a mistake made during her/his shift to the physician from the following shift, the latter might commit even larger mistakes. Like the heuristic to Commit potential collaborators to a course of action by closing their escape routes! it is easy to see how this heuristic, emphasizing the importance of trustworthiness, may guide, moreover, B2B-relations.

Insight – principle of cutting-through

“The fundamental goal of science is to find invariants” (Simon, 1990, p. 1). The invariants scientists aim to uncover are often studied within disciplinary silos – and this is also how many research institutions are organized: Physicist posit physical laws, management scholars those governing organizations, and so on. Yet, heuristics and their environments cut across disciplinary and professional boundaries. Hence, to discover mental and environmental invariants, one must look across disciplines. In this sense, the study of heuristics and environments needs to avoid the traps of disciplinary segregation and unnecessary professional specialization.

Based on the cutting-through principle, six meta-heuristics (No. 8–13) can be formulated. All operate on a common building-block – one Simon (1990) believed to represent one “of the mechanisms used by human bounded rationality to cope with real-life complexity” (p. 8): recognition. Looking across disciplines, professions and other silos helps – rather than looking within them – to recognize invariant patterns, namely both (1) how a given heuristic or its building-blocks are re-emerging or re-discoverable against the backlight of seemingly different contexts, and (2) how those seemingly different contexts exhibit common characteristics key to the functioning of that specific heuristic. Those patterns are the ‘invariants’ of the mental and the environmental blades that the science of heuristics aims to uncover, and that meta-heuristics 8–10 directly focus on.

Gigerenzer (2022) provides principles for running a research group, which operationalize meta-heuristics 8 and 9. For example, he suggests that research concentrates on problems as those appear in the real world, not as framed in particular disciplines. Katsikopoulos et al. (2024) describe how meta-heuristic-8, in bringing two researchers with different disciplinary backgrounds together, led to novel predictions concerning the less-is-more effects with the recognition heuristic in group decision-making. Hamlin (2017) did interdisciplinary research in line with meta-heuristic-10. He uncovered a key similarity of sports, biology and aviation, tracing how the simple interception rule, the gaze heuristic, aided the Royal Airforce’s triumph over the German Luftwaffe in the Second World War, and how it may be used by animals and athletes to catch prey and balls respectively.

Not just cutting through different disciplines can aid uncovering heuristics. Also using multiple means for inquiry can permit recognizing, and ultimately connecting the dots. Since its existence, the FFH program has embraced multiple methods, starting with computer simulations, mathematical analyses, experiments in the lab and in the wild. We added a few more in this article, including the study of proverbs (see also Atanasiu et al., 2023; Polya, 1945), ancient writings, folk-wisdom or even fairy tales. And those pathways complement other qualitative research that relied on interviews with managers and other practitioners to discover heuristics (e.g. Bingham and Eisenhardt, 2011). Indeed, we are practitioners ourselves – of the science of heuristics. And as other practitioners, in reporting on ‘our’ heuristics, we relied on introspection and self-observation. Hence, Know thyself! can be thought of as another meta-heuristic for discovery we used to write this article.

But there are yet other, fully unexpected paths to discovery, too. For example, a commander of the Royal Airforce discovered the gaze heuristic accidently prior to the Second World War, when trying to improve fighter interception capability (Hamlin, 2017). And the recognition heuristic was discovered incidentally when analyzing data of an experiment on overconfidence (Katsikopoulos et al., 2024). So, the path of serendipity is yet another possibility for discovering heuristics. Can one prepare for taking this path? Pasteur pointed out: “chance only favors the prepared minds” (Vantomme and Crassous, 2021, p. 597). To us, this suggests that one needs to be both open and ready to discover. We have experienced the benefits of openness and readiness in discovering the take-the-last heuristic for writing.

We can now tie all this in. Aged 100, former prosecutor at the Nuremberg Trials, Ferencz, published a book on lessons he had learned in his life. His book is a treasure box of heuristics. One of Ferencz’s lessons is: “It is possible to learn wherever one is. When you watch a film, read a book, walk in a street, have a conversation — do not be passive. Everything you do offers the opportunity to learn something new, and you never know when that knowledge will prove itself to be useful. (Ferencz, 2020, p. 45; Italics added, translated). This lesson told by Ferencz matches three meta-heuristics – No. 11–13 – we use (Table 2). As examples of meta-heuristics 12 and 13, consider yourself as you are reading this article. You might not agree with everything that was said. Nevertheless, we would like to think that, if you were open while reading, this article, though unconventional, opened doors and may even contribute to shaping-up new ideas.

Being open implies wanting to listen to others – notably to ideas that differ from one’s own. German Chancellor Schröder, allegedly included on purpose individuals in his government who would contradict him (Schäfer, 2023), and during the Cuban missile crisis, President Kennedy deliberately fostered skeptical attitudes in his policymaking group (Janis, 1982). “Be sure to include a contrarian” (p. 5) was the principle that Gigerenzer (2022) designed into the ABC-Research-Group when he set it up. Meta-heuristic-13 fosters openness and learning through contradiction. Indeed, as we witnessed, the science of FFH has benefitted enormously from scholars who disagreed with its propositions, and who in challenging those, shaped progress. For instance, it was the critique of Dougherty et al. (2008) and others that motivated methodological (Scheibehenne et al., 2013) and theoretical developments, including the ACT-R simulations referred to above (see Katsikopoulos et al., 2024; for more details). It was Bröder’s critical work on take-the-best that shaped theory on that heuristic and methods to infer its use from experimental data (e.g. Bröder and Schiffer, 2003).

To conclude, the principle of plurality comes with the principle of correspondence, and is reflected therein, namely in the diversity of the mappings between the environmental and the mental blades. Those mappings, in turn, are best understood as systems of connected elements, as suggested by the principle of connectedness. In line with those three principles, the principle of cutting-through suggests an interdisciplinary approach to studying heuristics and their environment. And one can see similar manifestations at a meta-level: the diversity of methods and fields, implied by the principle of cutting-through, reflects, at a meta-level, the principle of plurality, and the fit between methodological tools and the different types of targets of those tools (e.g. qualitative accounts of heuristics and their environments as opposed to quantitative ones), reflects, at a meta-level, the principle of correspondence.

4. Discussion

The principles and meta-heuristics formulated in the previous sections may be useful to aid overcoming the research-gaps we pointed to at the start of this article. For instance, we highlighted how the growing interest in heuristics in the management-literature (e.g. Atanasiu et al., 2023; de Treville et al. 2023; Guercini and Lechner, 2021; Hodgkinson et al., 2023; Katsikopoulos, 2023) lacks an integrative framework, which is problematic for scholars studying how heuristics are discovered. One may add that the changing technological and geopolitical environment – and the resulting uncertainties – may lead management-practitioners to perceive a need to equip themselves with new principles for behavior. We speculate that many of those principles may not be new but reflect ancient insights that aided humans to navigate past uncertainty-spreading “winds of change,” as they are captured, for example, by proverbs (Marewski and Hoffrage, 2020, p. 281).

4.1 How can one study the proposed heuristics empirically?

The FFH-program poses four types of questions (Gigerenzer et al., 2008). Descriptive questions ask what heuristics agents use in what environment. Ecological questions ask in what environment each heuristic performs well. Applied questions focus on how decision-making can be aided, say by improving the fit between people’s heuristics and their environments through teaching and policymaking. Methodological questions ask how descriptive, ecological and applied questions can be addressed.

Those questions could guide research on all heuristics, including our meta-heuristics and diverse (e.g. ancient, proverbial) others. Such research may, furthermore, be guided by the four unifying principles we formulated above.

4.2 Limitations

Our propositions come with several limitations. For instance, are our meta-heuristics just-so stories? Pointing out that certain research is consistent with a meta-heuristic – as we did in Section 3 – does not imply that the meta-heuristic was used. Moreover, our self-reports entail views that are primarily referable to the environments in which we have gained our experiences. Could it be that our propositions are too specific to our own contexts – implying that they might find little agreement even with other neo-Simonian researchers? Empirical research should tackle what one may also see as problem of scope; for example by turning to experiences made by other research-practitioners. All heuristics must be submitted to rigorous tests, asking, for example, to what extent and when a given heuristic aids decision-making more than other procedures.

Yet other issues come with the nature of the scissors: They come as multiples that correspond to, and that are connected with, each other, and that moreover, cut through disciplinary contexts. Interdisciplinary research is per se a difficult enterprise; however, as we have illustrated, heuristics and their environments can seemingly be found everywhere – from the military to writing. This makes not only conceptual clarity paramount but also intensive debate with neighboring work. For instance, prior to writing this article, we were not aware of the notion of meta-heuristics in computer science – it was a reviewer who alerted us to that. Likewise, we learned that we ought to consult research on analogies. The problem: we do not know what we do not know.

5. Conclusion: the Gestalt of Simon’s scissors

In studying the mutual dependence of minds and their environments, and the scissors they jointly produce, Simon took up a tremendous challenge. Figure 6 symbolizes that challenge. One can switch between seeing a vase or two faces. But it is very difficult to perceive both images simultaneously. Trying to make visible at the same time the mind and its environment, or equivalently the environment and its mind, is perhaps Simon’s (Figure 7) greatest legacy. We hope that this article can help spawn such work, particularly when it comes to discovering heuristics for management decisions.

Figures

Figure 1

The healthcare professional’s decision tree

Figure 2

Experiencing Simon’s scissor metaphor by writing with take-the-last

Figure 3

How the environment shapes complexity in behavior

Figure 4

What do you recognize between the branches of the two olive trees?

Figure 5

How the 5-Why-method bets on the last element encountered

Figure 6

The scissors are a Gestalt

Figure 7

Herbert Simon. Once Simon’s image here has been stored in recognition memory, you will see Simon again, looking at you through the olive trees shown in Figure 4

Table 1

Definition of heuristics and meta-heuristics

Definitions
Heuristics
A heuristic is a simple, actionable guiding-principle for behavior that reflects insights about the world. Those insights can be thought of as hypotheses about how to find good solutions in specific environments, characterized by uncertainty, as opposed to calculable risk. Simplicity is achieved by selecting information, reducing computation and relying on basic human capacities, mental or physical
Meta-heuristics
A meta-heuristic is a simple, actionable guiding-principle for discovering heuristics. Meta-heuristics reflect insights about heuristics. These insights can be thought of as hypotheses about how to find good heuristics in specific environments, characterized by uncertainty, as opposed to calculable risk. Simplicity is achieved by selecting information, reducing computation and relying on basic human capacities, mental or physical

Source(s): Authors’ elaboration

Table 2

Meta-heuristics

Number	Meta-heuristic
1	To study behavior, focus on describing a toolbox of heuristics and their respective environments, rather than on taking a one-size-fits all approach
2	Develop models that allow understanding which blades fit to each other, that is, when decision makers rely on which tool in their repertoire as a function of their environment
3	To describe environments in which a given heuristic is operating, spell out in as much detail as possible the heuristic, that is, precisely specify/define the heuristic under consideration
4	To describe heuristics which are successful in a given environment, spell out in as much detail as possible the environment, that is, precisely specify/define the environment under consideration
5	Engage in micro-level systems-thinking. To study a heuristic, investigate how it nestles into other components of cognition (e.g. memory, perception) as well as how the resulting cognitive system emerges by interacting with the external environment
6	Engage in macro-level systems-thinking. To study a heuristic, investigate how it interplays with the behavior of other actors
7	To study discovery, investigate how different heuristics for discovery interplay with each other and the structure of research environments
8	Those who want to study heuristics and environments need to collaborate with actors beyond their own disciplinary field. To do this, broad readings of heuristics and environments need to be developed that can cut across disciplines
9	When studying heuristics and environments, interview experienced practitioners working in different, seemingly unrelated professions. Across these professions, inquire to what extent the same heuristics are used for decision-making
10	To better understand a heuristic, write articles that try to trace it and its environment across different fields
11	As you are reading, running studies, analyzing models, or just going through daily life, note candidate heuristics that you come across on a list, to be used for future in-depth inquiry
12	In order to learn about heuristics and environments, be curious and open: Serendipity, systematic research, multiple sources and tools for inquiry all enable discovery
13	Search for, listen to – and ideally collaborate with – researchers who might disagree with you

Note(s): For further discussion of meta-heuristics 8, 12 and 13, see Katsikopoulos et al. (2024)

Source(s): Authors’ elaboration

Conflict of interest: The authors declare that they have no conflict of interest.