Solving the Entrepreneurial Puzzle: The Role of Entrepreneurial Interpretation in Opportunity Formation and Related Processes

Authors


Ilídio Barreto, Católica Lisbon School of Business and Economics, Universidade Católica Portuguesa, Palma de Cima, 1649-023 Lisboa, Portugal (iba@clsbe.lisboa.ucp.pt).

abstract

Opportunity formation plays a central role in the entrepreneurship literature. The two dominant perspectives on this topic (discovery view and creation view) tended to consider search and action as the main mechanisms. Drawing on strategic issue interpretation view and managerial cognition perspective, we argue for the inclusion of a third mechanism (entrepreneurial interpretation). Specifically, we develop the boundary assumptions and testable propositions of an entrepreneurial interpretation model. Then, we show how entrepreneurial interpretation informs both discovery and creation processes. Overall, our theory provides an expanded understanding of how individuals form and decide to exploit opportunities.

INTRODUCTION

The question of how individuals and organizations form and exploit opportunities has long been a central topic in the entrepreneurship literature (e.g. Kirzner, 1973; Schumpeter, 1934). Prior research has depicted opportunity formation in several ways, such as creation (e.g. Alvarez and Barney, 2007; Sarasvathy, 2001), discovery (e.g. Shane, 2003; Shane and Venkataraman, 2000), identification (e.g. Corbett, 2007; Ucbasaran et al., 2008), and recognition (e.g. Baron, 2007; Kirzner, 1973), among others. Such a proliferation has led to some confusion within the field due to the fact that the same concept has been used to characterize different processes, and different concepts have been used to characterize the same underlying process.

Recently, Alvarez and Barney (2007) provided a clarification of opportunity formation and exploitation processes by proposing a categorization based on the nature of the opportunity source. They suggest that a discovery view applies to processes where opportunities arise from exogenous shocks, such as technological, regulatory, political, social, or demographic changes (Kirzner, 1973; Shane, 2003). According to this perspective, opportunities already exist independently of the focal actor's efforts, waiting to be discovered (Shane and Venkataraman, 2000). In contrast, a creation view applies to processes where opportunities emerge from endogenous shocks, that is, they are socially constructed by the focal actor and other relevant peers through a chain of action and reaction (Weick, 1979). For the creation perspective, opportunities do not exist out there waiting to be discovered. Instead, they are created by the entrepreneurs' efforts (Baker and Nelson, 2005; Sarasvathy, 2001).

All research streams in this field soon recognized that entrepreneurial activities occur under conditions of uncertainty due to the amount and nature of information that is actually available (Kirzner, 1973; Schumpeter, 1934). Uncertainty runs along a continuum from complete ignorance to near certainty (McMullen and Shepherd, 2006). Before the occurrence of any kind of shock, complete ignorance is likely. In discovery processes, the entrepreneurial actor's situation before the shock is one of ‘(utterly unknown) ignorance’ (Kirzner, 1997, p. 72) or complete ignorance (McMullen and Shepherd, 2006; Miller, 2007). In creation processes, the focal actor is in a situation of ‘complete ignorance’ (Ellsberg, 1961) or ‘fundamental uncertainty’ (Dequech, 2006) since the relevant information, at the moment when the social construction begins, is not yet created (Alvarez and Barney, 2007).

The occurrence of a shock tends to transform the ignorance level of the entrepreneurial actor. In some occasions, the shock may make possible the identification of feasible outcomes and the estimation of their probabilities (Miller, 2007). This is why discovery studies have typically focused on contexts where the information generated by the exogenous shocks enables the estimation of the opportunity's expected value (Casson and Wadeson, 2007; Shane, 2003; Shane and Venkataraman, 2000). However, in many circumstances, the informational setting after either an exogenous or an endogenous shock may be characterized by equivocal and incomplete information (Daft and Lengel, 1986; Daft and Macintosh, 1981; Dutton et al., 1983; Ellsberg, 1961; von Hayek, 1945), rendering it hard, if not impossible, to identify all possible outcomes, let alone to estimate their probabilities (March and Shapira, 1987).

Thus, some interesting research questions emerge: First, what is the mechanism that entrepreneurial actors tend to use when facing equivocal and incomplete informational settings, and when does such a mechanism lead to important entrepreneurial decisions and actions (including decision to exploit)? Second, what are the specific implications of such a mechanism for discovery and creation views?

In this study, drawing on the strategic issue interpretation (Dutton and Jackson, 1987) and managerial cognition (Walsh, 1995) perspectives, we attempt to address those research questions by proposing an entrepreneurial interpretation theory. Our paper makes several contributions to the entrepreneurship literature. First, we extend past focus on search and action by introducing an overlooked but crucial mechanism (entrepreneurial interpretation) through the formulation of the assumptions and testable propositions of a novel theoretical model. Second, in response to the alleged fuzziness of the alertness concept (e.g. Alvarez and Barney, 2002), dominant in prior discovery literature, we propose new, clearly defined, mechanisms (automatic interpretation and directed search), and in addition we clarify the boundaries of existing ones (undirected search). Third, we expand past reliance on action and reaction within creation processes (Alvarez and Barney, 2007) to include in a systematic way a new mechanism (constructed interpretation) that helps explain how actors operate as the informational setting evolves from ‘complete ignorance’ to a situation characterized by both equivocal and incomplete information.

PERSPECTIVES ON OPPORTUNITY FORMATION

Opportunity formation has long been recognized as a central element in the entrepreneurial process (Kirzner, 1973; Schumpeter, 1934; Short et al., 2010), and the relevance of such a role has been reinforced in more recent literature (Shane and Venkataraman, 2000; Stevenson and Jarillo, 1990). Opportunity is a business project that seems potentially worth pursuing (Casson and Wadeson, 2007; Sarasvathy, 2001). Because entrepreneurial action depends on opportunity formation processes, the profusion (and consequent confusion) of proposed processes within the extant entrepreneurship literature intended to explain how opportunities are formed is not surprising. One important clarification was recently suggested by Alvarez and Barney (2007). They argued that two major perspectives of opportunity formation should be considered according to the relevant type of source: a discovery view for exogenous shocks, and a creation theory for endogenous shocks.

The Opportunity Discovery View

In the discovery theory, opportunities are typically assumed to arise from exogenous shocks that disrupt competitive equilibrium conditions. As Kirzner put it, ‘[w]ithout autonomous changes in [consumer] tastes, or in technological possibilities, or in the availability of resources, no one can have any interest in altering his plans . . . The market is in equilibrium’ (1973, p. 11). Several kinds of exogenous shocks can be considered, including technological, regulatory, political, social, and demographic changes (Shane, 2003, p. 23). Within the discovery theory, opportunities were initially characterized as potentially profitable situations that are ‘readily available’ but ‘hitherto overlooked’ (Kirzner, 1997, p. 72). They exist in the external world waiting to be discovered (e.g. Baron, 2007; Casson and Wadeson, 2007). The situations that originate opportunities are considered as real and objective phenomena irrespective of actors' efforts (e.g. Shane, 2003; Shane and Venkataraman, 2000), and are recognized as opportunities based on their expected value (Casson and Wadeson, 2007; Shane and Venkataraman, 2000): i.e. actors can anticipate possible outcomes associated with the opportunity as well as their probabilities (Bernoulli, 1954 [1738]; Knight, 1921; von Neumann and Morgenstern, 1944).

One key concept for the understanding of the mechanisms underlying the discovery theory is alertness. The importance of alertness was proposed by Israel Kirzner (1973, 1979, 1997, 1999, 2009) in a series of works on entrepreneurship. He considered alertness ‘the entrepreneurial element in the economic behavior’ (Kirzner, 1973, p. 15). Unfortunately, as we will discuss in greater detail later in this article, it is possible to discern several conceptualizations of alertness within this body of work, a situation that has contributed to some fuzziness in its use within the literature (Alvarez and Barney, 2002), and, in consequence, to some dispute about whether discovery is based on alertness or search (Shane, 2003, p. 264).

In one of his works, Kirzner defined alertness as ‘the ability to notice without search opportunities that have hitherto been overlooked’ (1979, p. 48; emphasis added). This conceptualization of alertness has been largely used in the discovery literature (e.g. Shane, 2003; Shane and Venkataraman, 2000). For instance, Shane (2000) showed that entrepreneurs can discover opportunities without searching as long as those opportunities are related to actors' prior experience and education. Later on, Kirzner (1997) attempted to position alertness-based discovery midway between Stigler's (1961) directed search, and pure luck. On the one hand, he attempted to distinguish discovery from directed search. In his classic work on the economics of information, Stigler suggested that the detection of profitable situations can result from information search, which people engage in as long as the marginal benefits of acquiring additional information exceed its marginal costs. The ‘Stiglerian’ search is termed directed search because it assumes that people know what they are looking for (i.e. the missing pieces of information) and have a task in mind to direct their search (e.g. Browne et al., 2007). By definition, Kirzner argues, discovery opportunities cannot be the object of directed search because in such cases ‘one is not aware at all that one has missed the grasping of any profit’ (1997, p. 71; emphasis added), so people do not know ‘what to look for’ (1997, p. 72). Thus, according to Kirzner, what distinguishes discovery from directed search is that ‘the former (unlike the latter) involves that surprise which accompanies the realization that one had overlooked something in fact readily available’ (1997, p. 72).

On the other hand, in response to criticisms of his earlier proposals (Demsetz, 1983), Kirzner also attempted to distinguish his alertness-based conceptualization of discovery from pure luck: ‘What accounts for a systematic tendency toward that succession of wholesome surprises . . . is not any implausible series of happy accidents, but rather the natural alertness’ (1997, p. 72). In addition, he specified the meaning and the importance of alertness by clarifying that ‘[w]ithout knowing what to look for . . . the entrepreneur is at all times scanning the horizon’ (1997, p. 72). This latter point is relevant as it suggests that Kirzner, at least in this conceptualization, also included some kind of search since he equated alertness with ‘scanning at all times’. Furthermore, this view is consistent with the early association between competition and discovery processes advanced by Friedrich von Hayek, whose works strongly influenced Kirzner's ideas, in his ‘Competition as a discovery procedure’ article: ‘competition is important primarily as a discovery procedure whereby entrepreneurs constantly search for unexploited opportunities that can also be taken advantage of by others’ (von Hayek, 2002 [1968], p. 18, emphasis added).

Given his remarks in relation to directed search, this second conceptualization of alertness might be properly viewed as undirected search. That is, discovered opportunities can be regarded as the result of undirected search – i.e. when people search for existing information that would lead them towards an opportunity, whereas they do not know exactly what they are looking for. An important body of research already exists within the discovery perspective suggesting the importance of different types of undirected search, such as broad scanning (e.g. Kaish and Gilad, 1991), general information search (e.g. Ucbasaran et al., 2008), active search (e.g. Ozgen and Baron, 2007), and systematic search (e.g. Fiet, 2007).

In sum, a significant body of research from the discovery perspective suggests that opportunities can arise either from processes without search or from undirected search. However, some important questions remain: What is the theoretical mechanism that explains the ‘discovery without search’ when the informational setting is characterized by both equivocal and incomplete information that precludes the estimation of the opportunity's expected value? If directed search can be considered as a valid mechanism for discovered opportunities, what guides such a search? Are there other relevant mechanisms not yet considered in the literature?

The Opportunity Creation View

Whereas in the discovery view opportunities are typically formed by exogenous shocks, in the creation view opportunities are endogenously constructed through entrepreneurs' actions and reactions, making possible their detachment from existing markets or industries (Alvarez and Barney, 2007). Instead of considering opportunities as the outcome of objective phenomena that exist before any intervention of the entrepreneur, the creation approach assumes that opportunities are social constructs that do not exist independent of entrepreneurs' perceptions and actions (Weick, 1979). So opportunities are created (rather than discovered) as part of the entrepreneurial process (Sarasvathy, 2001): actors create something from nothing (Baker and Nelson, 2005).

As the opportunity does not exist prior to the entrepreneur's actions, decision makers cannot anticipate the opportunity's possible outcomes or the probabilities associated with those outcomes (Alvarez and Barney, 2007): they are in a situation of ‘complete ignorance’ (Ellsberg, 1961; McMullen and Shepherd, 2006). The would-be entrepreneur starts the creation process ‘without any certainties about the existence of a market or a demand curve, let alone a market for his or her product, or a potential revenue curve’ (Sarasvathy, 2001, p. 249).

In addition, if creation opportunities do not exist before the actor's intervention, there is no point in searching for existing information that could lead to an opportunity because the relevant information is not there yet. Instead, the main mechanism within creation processes is enactment. Enactment is the process through which ‘managers construct, rearrange, single out, and demolish many “objective” features of their surroundings’ (Weick, 1979, p. 164). That is, instead of a process in which actors deal with an objective, exogenous environment, enactment means that ‘people often produce part of the environment they face’ (Weick, 1995, p. 30). The creation theory assumes that, whereas a deliberate or even random variation can trigger an opportunity formation process, the typical starting point is likely to be blind variation (Alvarez and Barney, 2007) or a generalized end goal (Sarasvathy, 2001). Blind variations are those ‘produced without prior knowledge of which ones, if any, will furnish a selectworthy encounter’ (Campbell, 1960, p. 381). They correspond to situations ‘when people are on their way to something else’ (Aldrich and Kenworthy, 1999, p. 20). In addition to such unconscious starting points, the formation process of creation opportunities can also be triggered by actors' generalized end goals; that is, some rather generic and abstract human aspirations pursued by would-be entrepreneurs (Sarasvathy, 2001).

According to the creation approach, it is important to note that at the moment when the formation process begins, actors typically do not have any idea whether they are (or not) on a path towards an opportunity (Aldrich and Kenworthy, 1999; Alvarez and Barney, 2007; Shah and Tripsas, 2007). As the process unfolds, a succession of actions, reactions, and perceptions follows. Such an enacted process is essentially iterative and social. On the one hand, opportunity creation is about action, experimentation, creativity, and playfulness (Aldrich and Kenworthy, 1999). Actors iteratively attempt to build ways to combine what they have at hand (Baker and Nelson, 2005), trying possible effects that can be created with a given set of means (Sarasvathy, 2001), testing conventional limitations (Baker and Nelson, 2005), or refusing the constraints of existing logics (Shah and Tripsas, 2007). Then, reaction succeeds to action. At each iteration, the entrepreneurs listen to their customers and other strategic partners' responses to the outputs of their actions, form their own perceptions, and readjust their generalized end goals (Sarasvathy, 2001). Then, new actions are initiated. So multiple variations are iteratively submitted to selection pressures, and, at each step, more relevant information for the opportunity formation is produced. On the other hand, the social nature of this enacted process is usually important. Opportunity emerges from intense interaction among involved actors due to the distribution of knowledge among different people (Baker and Nelson, 2005), and from the contingent interactions between the actions and perceptions of entrepreneurs and the aspirations of their partners (Sarasvathy, 2001). As a result, the enacted process tends to be highly path-dependent (Alvarez and Barney, 2007), and in consequence largely idiosyncratic (Baker and Nelson, 2005): even if two actors share the same starting point, the resulting creation opportunity would tend to be significantly different according to the successive contingencies developed within the process.

Whereas important developments have been already made in this emerging view, some crucial questions remain: Why are cognitive representations important and how exactly do they work in this perspective? How and when does each component of the creation process lead to another component? When does the process stop?

ENTREPRENEURIAL INTERPRETATION THEORY

In the previous section, we showed that there are two main perspectives on opportunity formation: a discovery view for opportunities generated by exogenous shocks, and a creation view for opportunities generated by endogenous shocks. We concluded that several important questions remain unaddressed by these two approaches. In this section, we develop an entrepreneurial interpretation theory that is mainly based on the strategic issue interpretation literature, and explain why and how it is particularly suitable to address such questions.

The Strategic Issue Interpretation Literature

The strategic issue interpretation literature aims to explain how managers learn about their external environments and implement corresponding actions. Specifically, this view argues that managers cope with strategic issues through interpretation (Dutton and Jackson, 1987). Strategic issues are defined as trends, events, and developments that can have a major and discontinuous impact on the firm's present and future actions (Dutton et al., 1983). Issues of relevance within this stream of research are those associated with ‘conflicting’ and ‘insufficient’ information (e.g. Dutton et al., 1983; Dutton and Jackson, 1987). Interpretation is considered as ‘the process by which managers translate data into knowledge and understanding about environment’ (Daft and Weick, 1984, p. 291).

Action in response to these issues requires giving meaning to them. Managers give meaning to data related to a specific issue through the imposition (upon that data) of schemas or knowledge structures (Dutton and Jackson, 1987), which are mental templates of organized knowledge related to an information domain (Walsh, 1995). One salient and commonly used schema regarding strategic issues is ‘opportunity’ (Jackson and Dutton, 1988). A stronger activation of the opportunity template should be expected when the characteristics of a particular issue fit the actor's template well (Dutton and Jackson, 1987).

Substantial empirical evidence provides support for the strategic issue interpretation view. The determination of the contextual conditions, the ‘strategic issues’, is central to the study of opportunity interpretation: extant studies considered multiple events (Chattopadhyay et al., 2001; Martins and Kambil, 1999), single, generic situations (Julian and Ofori-Dankwa, 2008), single, specific, and discontinuous shifts (Gilbert, 2006), and single, specific, continuous situations (Denison et al., 1996; Sharma, 2000; Thomas and McDaniel, 1990; Thomas et al., 1993). Studies have examined variation across firms in a single industry (e.g. Chattopadhyay et al., 2001; Martins and Kambil, 1999; Sharma, 2000; Thomas and McDaniel, 1990; Thomas et al., 1993) or in multiple industries (e.g. Denison et al., 1996; Julian and Ofori-Dankwa, 2008). Research has tended to focused on perceptions of CEOs (e.g. Denison et al., 1996; Julian and Ofori-Dankwa, 2008; Thomas and McDaniel, 1990; Thomas et al., 1993), of top management team members (e.g. Chattopadhyay et al., 2001; Martins and Kambil, 1999), or of mixed sets of top executives and other managers (e.g. Gilbert, 2006; Sharma, 2000).

This body of empirical research showed that opportunity interpretation can be highly consequential for subsequent stages of decision-making, and action (e.g. Anderson and Nichols, 2007; Denison et al., 1996; Julian and Ofori-Dankwa, 2008; Kuvaas, 2002; Sharma, 2000; Thomas and McDaniel, 1990; Thomas et al., 1993). These studies, which typically undertake a firm-level analysis, provide ample evidence for the importance of opportunity interpretation. Moreover, they show that the effect of opportunity interpretation on action depends on the magnitude of such interpretation: high ratings on opportunity interpretation (that is, when the actor's opportunity template fits the focal issue well) tend to be associated with subsequent action, while low ratings (that is, when the actor's opportunity template fits the focal issue poorly) do not seem do induce action. Interestingly, one study that used a dichotomous measure for opportunity interpretation (rather than a continuous measure that would capture the varying magnitudes of the interpretation) did not find a significant effect of opportunity interpretation on action (Chattopadhyay et al., 2001). Its authors rightly acknowledged that this finding might be due to their ‘inability to discriminate between opportunities of varying magnitude’ (2001, p. 951).

Towards an Entrepreneurial Interpretation Theory

The strategic issue interpretation literature and the underlying cognitive theories provide a sound and proper foundation for an entrepreneurial interpretation theory. We begin by formalizing its key assumptions (about the source of opportunities, the informational settings, and the main mechanism involved).

First, we assume that both exogenous and endogenous (informational) shocks can constitute sources of opportunities. These shocks are emerging developments, trends, and events associated with significant informational shifts that can have a major and discontinuous impact on what individuals or organizations can do (Daft and Weick, 1984; Dutton, 1993a). Regulatory, technological, political, demographic, or social changes are examples of exogenous shocks. Endogenous shocks are developments actively constructed by actors themselves (Dutton, 1993b).

Second, we assume that only the shocks that are actually attended by individuals, groups, or organizations are subject to interpretation. Many trends and events might occur around actors. However, actors can only attend to some of them, ignoring all others (Daft and Weick, 1984).

Third, following the strategic issue interpretation literature, we focus on ambiguous situations characterized by ‘conflicting’ and ‘insufficient’ information (e.g. Dutton and Jackson, 1987; Dutton et al., 1983). Those situations are important and have been emphasized in quite similar ways by different streams of decision making research. In his seminal work, Ellsberg characterized them as ‘situations where available information is scanty or obviously unreliable or highly conflicting’ (Ellsberg, 1961, pp. 660–1). Friedrich von Hayek recognized that a major element of the problem of a rational economic order was that the required information and knowledge ‘never exists in concentrated or integrated form, but solely as the dispersed bits of incomplete and frequently contradictory knowledge which all the separate individuals possess’ (von Hayek, 1945, p. 519; emphasis added).

Thus, equivocality and missing information are two components of interest in our settings. Equivocality corresponds to the existence of multiple and conflicting meanings in the informational setting (Daft and Lengel, 1986; Daft and Macintosh, 1981), and usually implies confusion and lack of understanding (Daft and Lengel, 1986). Missing (or incomplete) information corresponds to the difference between the information required to perform a task and the information actually possessed by actors (Galbraith, 1977). Decision making studies showed that two distinct kinds of missing information can be considered: one, in which the missing information is unknown to the decision maker but could be known to others (e.g. Camerer and Weber, 1992); another, in which the missing information is unknown to anyone (e.g. Chow and Sarin, 2002). We term the former unknown missing information, and the latter unknowable missing information.

The fourth assumption is that the main mechanism used to cope with these shocks is interpretation. Interpretation refers to giving meaning to information originating from a given shock through the application of an opportunity template (or knowledge structure) to the amalgam of available data (Dutton and Jackson, 1987). The adoption of interpretation is explained by cognitive theories (that underlie strategic issue interpretation view) through the concepts of simplification and knowledge structures. Cognitive simplification processes help individuals to overcome the difficulties raised by contexts that involve equivocal and missing information by enabling the information processing required to decision making (Schwenk, 1984). The use of knowledge structures, which are mental templates that individuals impose on an information environment to give it meaning, represents a key cognitive simplification instrument (Walsh, 1995). Next, we explain how this mechanism works.

The use of an opportunity template involves two main tasks (as illustrated in Figure 1): selective attention and gap filling. Selective attention refers to noticing the pieces of information generated by the focal shock that are consistent with the actor's opportunity template and, at the same time, disregarding those pieces of information that are inconsistent with the opportunity template or are considered irrelevant for such purpose (Dutton and Jackson, 1987). Gap filling refers to a process whereby, when the information associated with the particular shock is insufficient, the resulting gaps will tend to be filled with information consistent with the actor's opportunity template (Dutton and Jackson, 1987). In fact, the activation of interpretation resembles the solving of a puzzle with both missing and redundant pieces, as Figure 1 illustrates.

Figure 1.

How opportunity interpretation works

Consider a given actor's opportunity template (Figure 1a). In the mind of this particular actor, an opportunity template is formed by five information elements here represented by the pieces A, B, C, D, and E. On the other hand, suppose that a given shock emerges and produces a certain amount of information (Figure 1b). This information set comprises information elements A, C, D, F, G, and H, which the actor suddenly faces in a disorderly manner. The would-be entrepreneur spontaneously attempts to impose her opportunity template upon the shock's information set. She acknowledges and puts together the pieces A, C, and D, which are the ones consistent with her opportunity template. At the very same time, she puts aside the pieces of information that are inconsistent with the template (pieces F, G, and H). The selective attention task is completed. Finally, the actor performs the gap-filling task: she assumes that pieces B and E might have been presented in the environment despite her failure to notice them (Figure 1c). The interpreted puzzle is completed. An opportunity is formed.

An Entrepreneurial Interpretation Effects Model

Based on the assumptions just outlined, we now provide a model of the entrepreneurial interpretation effects. Since entrepreneurship is about action (McMullen and Shepherd, 2006), it is important to delineate the predicted effects of opportunity interpretation on entrepreneurial decision-making and actions. Opportunity exploitation refers to building full-scale operations for new products or services derived from a business opportunity (Choi et al., 2008). Like opportunity formation, decision to exploit constitutes a core topic in entrepreneurship (Shane, 2003; Stevenson and Jarillo, 1990). Surprisingly, decision to exploit is among the most under-researched areas in entrepreneurship (Choi and Shepherd, 2004; Shook et al. 2003). Next, we develop a model of opportunity interpretation effects with a particular emphasis on the explanation of the decision to exploit and other related entrepreneurial decisions and actions. Figure 2 depicts the proposed conceptual model.

Figure 2.

A model of entrepreneurial interpretation effects

As discussed above, individuals may deal with shocks through interpretation. Specifically, they try to impose their opportunity templates on the information set associated with each specific shock. According to each situation, different levels of matching between the actor's opportunity template, and the information related to each shock should be expected. We term such levels of matching as interpretive resolution levels since they represent the extent by which the actor's opportunity template resolves the situation's equivocality and lack of information, and how clear the focal situation resembles an opportunity in actors' minds. High (moderate/low) levels of interpretive resolution should mean high (moderate/low) levels of actor's confidence on his or her opportunity interpretation of the specific shock.

We argue that opportunity interpretation is likely to lead to a decision to exploit only when the interpretive resolution levels are high, i.e. when the information set related with the focal shock clearly fits the actor's opportunity template. This expectation is consistent with some considerations previously made by Knight (1921) and Weick (1995). Knight suggested that ‘[t]he action which follows upon an opinion depends as much upon the amount of confidence in that opinion as it does upon the favorableness of the opinion itself’ (1921, p. 227). Similarly, Weick argued that whether action is more (or less) likely will depend not only on the content of beliefs about the future, but also on ‘the intensity with which those beliefs are held’ (1995, p. 95). Likewise, we argue that, in an interpretation situation, what matters is not only whether the situation fits somehow an actor's opportunity template but, most importantly, how good the fit is.

Since opportunity exploitation is likely to involve a substantial resource commitment by the focal actor, a decision to exploit should depend on whether the associated end state is considered desirable and feasible in the actor's mind (McMullen and Shepherd, 2006). Framing a shock as an opportunity is likely to result from and incorporate the positive ‘charge’ or emotion, and the sense of controllability necessary to induce subsequent action (Dutton, 1993b). However, as suggested by Choi et al., ‘entrepreneurs need to wait until their level of ignorance reaches an acceptable level before deciding on shifting focus to exploitation. This level is “acceptable” when it provides the entrepreneur sufficient confidence to proceed with the investments required for exploitation of the opportunity’ (Choi et al., 2008, p. 338). Interpretation, by definition, provides an understanding about a shock (Daft and Weick, 1984), and, in consequence, tends to reduce the level of ignorance perceived by actors.

When dealing with shocks, individuals are likely to have the sufficient positive ‘charge’ or emotion, and the sense of controllability to surpass the acceptable level of ignorance, and, in consequence, to pass from opportunity formation to opportunity exploitation, only when their interpretive resolution level is high. A high interpretive resolution level for a given shock means that the actor was able to deal in a highly satisfactory manner with both the equivocality and lack of information associated with the shock when imposing his or her opportunity template on the shock's information set. In other words, a high interpretive resolution level might correspond to a reduction of the actor's sense of ignorance to an ‘acceptable’ level. In contrast, if the interpretive resolution is low, the sense of remaining ignorance about opportunity components of the focal situation should be quite high, and thus individuals will lack the confidence to decide upon an exploitative action related to the emerging shock. Therefore, we posit the following:

Proposition 1: In the presence of a shock, opportunity interpretation is more likely to lead to a decision to exploit when the interpretive resolution is high than when the interpretive resolution is low or moderate.

Proposition 2: In the presence of a shock, opportunity interpretation is more likely to lead to inaction when the interpretive resolution is low than when the interpretive resolution is high or moderate.

At this point, one might wonder what happens when the actor's interpretive resolution is moderate. In such cases, on the one hand, the level of matching between the actor's opportunity template, and the information related to each shock might not reduce the ignorance to a level that is low enough to induce opportunity exploitation. On the other hand, it might be significant enough to be associated with some degree of confidence (and curiosity) that can generate something more than mere inaction in relation to the focal shock. We argue that moderate interpretive resolution levels might trigger some intermediate entrepreneurial actions depending on the nature of the main cause of ex post ignorance and (in some cases) the nature of the missing information.

Ex post ignorance refers to the level of ignorance experienced by focal actors after the activation of their interpretive mechanisms. A moderate interpretive resolution for a given shock means that the actor was not able to achieve a level of ex post ignorance sufficiently low enough to induce opportunity exploitation afterwards. That is, the focal actor was not yet able to deal, in a satisfactory manner, with the equivocal and missing information associated with the shock when imposing his or her opportunity template to the shock's information set. This substantial ex post ignorance might be mainly driven by the extent of equivocality that persists after the interpretation, or alternatively, by the extent of missing information that was not satisfactorily dealt with through the activation of the actor's interpretive mechanisms.

When the ex post ignorance associated with moderate interpretive resolution levels is mainly prompted by equivocality, the solution is not to gather more information. In equivocal situations, more information could mean adding more contradictory views, and in consequence could increase (rather than reduce) confusion and equivocality (Daft and Lengel, 1986). Instead, in such situations actors may try to engage in interpretive absorption actions to complement or even refine their insufficient interpretation. Interpretive absorption actions refer to activities by which focal actors, rather than obtaining more information, deliberately attempt to acknowledge and eventually share others' knowledge structures that could enrich their reinterpretation of the shock. For example, actors may try to exchange their interpretations with their peers through rich media, such as face-to-face group meetings (Daft and Lengel, 1986). At an intra-organizational level, others' interpretations might be more easily looked for when they are likely to reinforce focal actors' a priori political interests (Dutton et al., 1983). But interpretive absorption actions might also occur beyond a focal actor's organizational boundaries. For example, in ambiguous situations bankers may attempt to acknowledge and share the knowledge structures of key people in rival banks (McNamara et al., 2002). We expect that such interpretive absorption actions should be more likely when the interpretive resolution is moderate (i.e. when the actor's own interpretation is not strong enough to directly trigger a decision to exploit) and the implied level of ex post ignorance is predominantly caused by unresolved equivocality (i.e. when gathering more information is not particularly helpful).

Proposition 3: In the presence of a shock, opportunity interpretation is more likely to lead to interpretive absorption actions when the interpretive resolution is moderate and ex post ignorance is mainly driven by equivocal information than otherwise.

When missing information (rather than equivocality) dominates the causes of ex post ignorance in moderate interpretive resolution situations, however, other kinds of entrepreneurial actions can be expected. In such cases, the imposition of an actor's opportunity template on the focal shock's information set may remove an important part of the equivocality involved while not fully resolving the interpretation problem caused by the lack of information. Which actions will be likely to subsequently occur will depend, we argue, on the nature of the missing information.

We highlighted above the importance of distinguishing unknown missing information from unknowable missing information, where the former refers to missing information that is unknown to the decision maker but could be known to others, and the latter refers to missing information that cannot be known by anyone.

When the level of interpretive resolution is moderate and ex post ignorance is perceived as mainly caused by unknown missing information, we expect focal actors to implement actions to gather the information they lack. Such information gathering activities are guided by the result of prior imposition of actors' opportunity template on the emerging situation. That is, the interpretation mechanism allows the determination of which information is missing, and therefore which information should be looked for (Figure 1c). The reader should recall from our analysis of the discovery theory that this process corresponds to directed search (Stigler, 1961) in which individuals know what they are looking for, and sharply contrasts with undirected search process in which individuals do not know what to look for (Kirzner, 1997). For example, before making the decision to enter in Formula One competition, the businessman Vijay Mallya seems to have interpreted the withdrawal of Jordan Grand Prix from Formula One in 2007 as an opportunity (based on his previous experience on the periphery of the sport as a sponsor of Toyota and Benetton). But as he was not fully convinced, he first engaged in some specific information gathering that subsequently confirmed his early interpretation (Allen, 2008). Accordingly, we expect the following:

Proposition 4: In the presence of a shock, opportunity interpretation is more likely to lead to directed search when the interpretive resolution is moderate and ex post ignorance is mainly driven by unknown missing information than otherwise.

However, when the level of interpretive resolution is moderate and ex post ignorance is perceived as mainly caused by unknowable missing information, we expect entrepreneurial actions other than directed search. In such circumstances, there is no point in searching for information (even in a guided manner) because focal actors perceive the relevant information as not existing yet (Alvarez and Barney, 2007). On the contrary, those situations seem to require the production of new information, i.e. action rather than search. However, here the expected action is no longer the result of some myopic variations, as might happen in some situations. Instead, action will be guided by the results of prior interpretation. Therefore, we term such an oriented process as directed action.

That is, the imposition of actors' opportunity templates on the new situation allows the determination of which relevant information is missing (Figure 1c), and therefore which information needs to be produced through action. This process is consistent with the idea of opportunity construction suggested by Dutton (1993b). She proposed that opportunities might be constructed by a balancing act between passive perceptions and active development, a process that contrasts with ‘mere’ interpretation. Examples of directed action have been already offered in the entrepreneurship literature. For instance, 3M scientists might use their 15 per cent of discretionary time and resources in ‘working to solve particular puzzles and problems, and testing various hunches and hypotheses’ (Adner and Levinthal, 2008, p. 44). We argue that those hunches and hypotheses can be generated by early interpretations that, while not yielding the interpretive resolution level strong enough to pass from opportunity formation to opportunity exploitation, might still provide the required guidance for the actions to solve the puzzle when the main cause of the ex post ignorance is unknowable missing information.

Proposition 5: In the presence of a shock, opportunity interpretation is more likely to lead to directed action when the interpretive resolution is moderate and ex post ignorance is mainly driven by unknowable missing information than otherwise.

EXPANDING DISCOVERY AND CREATION VIEWS

The entrepreneurial interpretation theory enables an expanded understanding of the two dominant perspectives on opportunity formation and exploitation. In fact, our theory directly addresses the questions that, according to our review of discovery and creation perspectives, require further consideration. Next, we explain such implications starting with the questions pertaining to the discovery view.

What is the theoretical mechanism that explains the ‘discovery without search’ when the informational setting is characterized by both equivocal and incomplete information that precludes the estimation of the opportunity's expected value?  The discovery view assumes that opportunities emerge from exogenous shocks. When the information generated by the focal shock is rich enough to enable actors to identify all possible outcomes and estimate the respective probabilities, the mechanism in use is likely to be identification (e.g. Ucbasaran et al., 2008): actors will able to compute the expected value of the main implications of the shock and identify whether the shock is (or is not) an opportunity (e.g. Casson and Wadeson, 2007; Shane and Venkataraman, 2000). However, in many situations, equivocality and lack of information will prevail (e.g. Daft and Lengel, 1986; Daft and Macintosh, 1981; Dutton and Jackson, 1987; Dutton et al., 1983; von Hayek, 1945), precluding the identification of all possible outcomes and the estimation of their probabilities (March and Shapira, 1987). On such occasions, interpretation should be the mechanism at work. Interpretation enables actors to give meaning to a situation characterized by the existence of missing and equivocal information. Actors interpret the shock through the imposition of their opportunity templates on available information. Our entrepreneurial interpretation model predicts that actors are likely to make a decision to exploit when the emerging situation strongly resembles their opportunity template, that is, when the interpretive resolution is high. In contrast, when the interpretive resolution is low, our model predicts that actors are likely to opt for inaction.

In discovery processes, a specific mode of interpretation, automatic interpretation, is likely to be relevant. Automatic interpretation refers to the mindless or automatic imposition of an opportunity template on the information set generated by a given shock, involving ‘limited cognitive effort or expenditure of attentional and analytic resources’ (Dutton, 1993a, p. 341). This mode of interpretation is thus characterized by being discrete rather than continuous, spontaneous, and involuntary rather than effortful and intentional. So interpretation in the discovery perspective should correspond to a typically passive (rather than active) process, where actors behave as external observers that simply perceive what is going on ‘out there’. Consistent with the typical level of analysis used in both theoretical (e.g. Shane, 2003; Shane and Venkataraman, 2000) and empirical (Corbett, 2007; Shane, 2000) studies in the discovery tradition, automatic interpretation focuses primarily on individual-level processes (Dutton, 1993a). This focus implies that the relevant interpretive resolution level reflects the degree of confidence of individual actors (rather than groups of actors), and that crucial for the outcomes of these processes are individual (rather than collective) knowledge structures.

If directed search can be considered a valid mechanism for opportunity discovery, what guides such a search? Are there other relevant mechanisms not yet considered in the literature?  In ‘Kirznerian’ contexts, directed search make no sense since actors do not know what to look for (Kirzner, 1997). However, some authors following the discovery perspective acknowledged that directed search might also be considered (Shane, 2000). Yet, what guides such a search? Our model provides one relevant answer. In some circumstances, the imposition of an actor's opportunity template on the information set produced by the exogenous shock might resolve most of the involved equivocality but not the lack of information. In those circumstances, if ex post ignorance is perceived as mainly driven by unknown missing information (that is, information that is not known by the actor but that can be known by others), then our model predicts that direct search is likely to be triggered: actors know what they need to look for.

Two additional components of an expanded discovery view are uncovered by our theory: directed action and interpretive absorption actions. We expect directed action when the interpretation of a shock removes most of the inherent equivocality but not the missing information component, and the missing information is perceived as unknowable. On these occasions, since the information that actors lack seems to be unknown not only to them but also to others, the mechanism in use should be directed action, that is, the missing information should be produced rather than searched. However, in other situations, the ex post ignorance might be dominated by the remaining equivocality. In these situations, there is no point in obtaining more information. More information might only increase confusion. Our model predicts that in those situations interpretive absorption actions are likely: this kind of action might enable actors to acknowledge or share the knowledge structures of some of their peers.

Our entrepreneurial interpretation theory also enables an expanded understanding of the creation perspective. As we noted above, several crucial questions remain overlooked within this perspective. Next, we explain how the entrepreneurial interpretation theory addresses such questions.

Why are cognitive representations important and how do they exactly work in this perspective?  According to the creation view, whereas a deliberate or even random variation can trigger an opportunity formation process, the typical starting point is likely to be a blind variation (Alvarez and Barney, 2007). In other words, at the beginning of this process, actors are in complete ignorance of where initial actions will drive them. One crucial aspect is that as the process unfolds more and more information is generated by those actions, and thus the informational setting might evolve from complete ignorance to a situation characterized by both equivocal and missing information. At this stage, interpretation is required to cope with the situation since interpretation may translate incomplete and equivocal data into an understanding of the environment. In this creation process, a particular mode of interpretation is likely to be at work. Following Dutton (1993b), we called it constructed interpretation as opposed to automatic interpretation. Like automatic interpretation, constructed (opportunity) interpretation also involves the imposition of an opportunity template on available data in order to give meaning to such data (Dutton, 1993b). By providing meaning to existing situations, interpretation enables and guides subsequent decision-making and entrepreneurial actions. However, constructed interpretation differs from automatic interpretation in several important dimensions. Consistent with the view of creation opportunities as social constructs, it corresponds to a continuous-efforts process rather than a one-shot phenomenon. Here, interpretation is a component of a dynamic, iterative process where opportunities are gradually constructed through actors' efforts to achieve something rather than passively perceiving something that already exists. Thus, individuals in the constructed interpretation behave as engaged, committed actors rather than as external observers. Furthermore, the social nature of a creation process means that the analysis focuses primarily on the collective formed by the involved actors (and their social interactions) rather than on the individual actor.

Specifically, constructed interpretation is based on an evolving interaction of three major components: action, individual interpretation, and shared understandings. Action precedes interpretation: it is myopic or blind action that generates a significant flow of information which, at some point in time, triggers individual interpretations by focal actors. Individual actors interpret the current stock of information produced by their actions through the imposition of their opportunity templates. However, since in a creation view opportunities are socially constructed, it is the collective (opportunity) interpretation that matters. As noted by Daft and Weick, the distinctive feature of a collective interpretation is sharing: ‘a piece of data, a perception, a cognitive map is shared among managers who constitute the interpretation system’ (Daft and Weick, 1984, p. 285). Shared understandings or collective interpretations, rather than individual perceptions, are likely to drive subsequent group or organizational decisions and actions (Lyles and Schwenk, 1992) in the creation process. Those shared understandings are social constructed and rely on consensus or agreement (Lyles and Schwenk, 1992).

How and when does each component of the creation process lead to another component? When does the process stop?  Our entrepreneurial interpretation model helps to explain the dynamics of the creation process. We showed above how the first component of such a process, i.e. initial action, leads to both individual and collective interpretation (or shared understandings). Now, we highlight how and when the collective interpretation drives subsequent decisions and actions. In an individual-based approach, such as discovery, interpretation leads to the decision to exploit when the interpretive resolution level is high, and, in such a case, that means that the actor's confidence on her opportunity interpretation is strong. In a group-based approach, such as creation, the relevant interpretive resolution level is the one associated with the collective interpretation, and depends on the degree of consensus or agreement among involved actors. In this entrepreneurial mode, a high interpretive resolution would mean that actors were able to develop a high degree of agreement regarding the fit between the situation and the actors' opportunity templates. If collective interpretation reaches such high interpretive resolution, then a decision to exploit is likely. Conversely, a low interpretive resolution would mean that actors converge on a high degree of agreement that the situation fits poorly (or does not fit at all) the actors' opportunity templates. If that is the case, a decision to abandon the ongoing project is likely. The remaining possible cases fall in the moderate interpretive resolution category.

In these remaining cases, as we explained in our entrepreneurial interpretation model, the dominant cause of ex post ignorance may vary. If the dominant cause is equivocal information, then collective interpretation should lead to interpretive absorption actions. If the dominant cause is unknown missing information, then directed search will follow the collective interpretation. If the dominant cause is unknowable missing information, then direct action is likely to be triggered. It is noteworthy that our interpretive absorption actions and directed search should be seen as corresponding to what past research labelled as the ‘reaction’ step in the chain process of creation. So, depending on the dominant cause of ex post ignorance, focal actors might deliberately attempt to acknowledge and eventually share the interpretations that their external peers make of the situation (to reduce the equivocality problem), or deliberately try to collect the specific and relevant information that they know is missing (to reduce the missing information problem). Finally, our prediction of directed action emphasizes the fact that not only are the successive rounds of action more and more focused (guided by the emerging interpretations) but also, and perhaps more importantly, that such actions are likely to occur only when actors perceive the missing information as being unknown for them as well as for anybody else (rather than being an universally used instrument).

In sum, our entrepreneurial interpretation theory not only provides expanded approaches for discovery and creation perspectives, but also highlights further important differences between them. Table I summarizes such additional differences.

Table I.  Expanding discovery vs. creation differences
 The discovery viewThe creation view
Interpretive mechanismAutomatic interpretationConstructed interpretation
Primary level of analysisIndividual levelGroup level
Nature of the interpretive process (I)DiscreteContinuous
Nature of the interpretive process (II)PassiveActive
Nature of actors' behaviourActors behave as external observersActors behave as engaged players
Main driver of interpretationIndividual knowledge structuresShared knowledge structures

DISCUSSION

Our aim in this paper was to provide an expanded understanding of how individuals form opportunities, when the involved opportunity formation mechanism is likely to lead to important entrepreneurial decisions and actions, and from which implications for discovery and creation perspectives can be derived. To this end, we reviewed past literatures on discovery and creation, and highlighted some crucial questions that remained unanswered. We then developed an entrepreneurial interpretation theory by explaining its main mechanism, setting its boundary assumptions, and formulating testable propositions; and we finally elaborated on the specific ways this theory expands discovery and creation perspectives.

The contributions made by this study shed light on a central topic in the entrepreneurship literature, opportunity formation and exploitation (Short et al., 2010), and fill major gaps within the two dominant views (discovery and creation). We now explain the overall contribution to the entrepreneurship field, and then discuss the contributions of our theory to each of the currently dominant views.

This paper contributes to the entrepreneurship literature by introducing an often overlooked mechanism, entrepreneurial interpretation, based on a systematic and coherent theoretical framework. While extant research focused on mechanisms such as identification (e.g. Ucbasaran et al., 2008), undirected search (e.g. Kaish and Gilad, 1991), or action and reaction (e.g. Alvarez and Barney, 2007), we expand this body of studies by showing that the relevant mechanism at work can be entrepreneurial interpretation: the way actors give meaning to available information through the imposition of an opportunity template on that information set. We further argue that this mechanism is particularly relevant in ambiguous informational settings characterized by both equivocal and incomplete information. Such informational settings are quite distinct from the situations that attracted most of the attention of past research, namely, the situations of complete ignorance, or situations of limited ignorance where the estimation of the expected value of the opportunity is possible.

This interpretive mechanism, we argued, is particularly powerful in explaining entrepreneurial decision-making and action, namely, decision to exploit. The surprising lack of research on decision to exploit has been acknowledged in the field (e.g. Shook et al., 2003). One notable exception is Choi et al.'s (2008) paper. They suggested that individuals decide to engage in opportunity exploitation only when their level of ignorance is reduced to an ‘acceptable’ level. Our model expands their explanation to settings where information is both equivocal and incomplete. We show that, in such contexts, ignorance reduction is achieved through interpretation, and individuals are likely to pass from opportunity formation to opportunity exploitation only when their interpretive resolution level is high; otherwise, they will not hold enough positive ‘charge’ or emotion and the sense of controllability necessary to surpass the acceptable level of ignorance, and substantial doubts will remain.

Our model also suggests that past focus on information asymmetries to explain different entrepreneurial behaviour among different actors (e.g. Shane, 2000) should be complemented by an emphasis on interpretive asymmetries: different entrepreneurial outcomes might be not only the result of people possessing different information, but also the result of people applying different opportunity templates even to the same information set. So this paper opens new avenues for future research as it suggests that systematic patterns of interpretation across individuals or groups of individuals should occur, influencing subsequent entrepreneurial actions. Future studies should also attempt to advance our knowledge in the field by discriminating the effects of information asymmetries from the effects of interpretive asymmetries. In addition, the study of the antecedents of entrepreneurial behaviour should benefit from an explicit consideration of entrepreneurial interpretation as a mediating factor. For instance, empirical evidence that showed that prior experience and education can influence opportunity formation (Shane, 2000) might be explained by the effect that prior experience and education have on opportunity interpretation: those are the typical factors that are expected to influence actors' opportunity templates or knowledge structures (Walsh, 1995).

Implications for the Discovery View

Our study has crucial implications for the discovery view. On the one hand, we introduce a specific mechanism, automatic interpretation, which has received little attention so far. Some confusion regarding the relevant decision making context seems to prevail in this literature. Notable works often treat risky and uncertain contexts interchangeably (e.g. Kirzner, 1997; Shane, 2003). However, other important studies emphasized that opportunities are identified through the estimation of the opportunity's expected value (e.g. Casson and Wadeson, 2007; Shane and Venkataraman, 2000), which involves the assumption that the identification of feasible outcomes and the estimation of their probabilities are possible. It is possible that the typical source of opportunities in a discovery view, i.e. exogenous shocks, produces information whose amount and quality are consistent with such a context. Yet, a large body of work in the literature argues and shows that many if not most exogenous shocks are likely to be associated with equivocal and largely incomplete information (e.g. Dutton et al., 1983). In such cases, a mechanism other than identification is likely to be at work as expected value cannot be estimated in such conditions (March and Shapira, 1987), but no systematic effort has been made so far to highlight such a mechanism. We address this gap by showing that automatic interpretation is the mechanism likely to be used when exogenous shocks produce informational settings characterized by equivocal and missing information. Automatic interpretation is a specific mode of entrepreneurial interpretation that refers to the mindless imposition of an opportunity template on available information with limited cognitive or analytical effort. Consistent with the discovery tradition, it corresponds to a spontaneous and passive (rather than an effortful and active) process where focal actors behave as external observers.

On the other hand, our entrepreneurial interpretation theory provides the infrastructure for a major clarification of the mechanisms previously suggested in the discovery literature as well as a reevaluation of the relevance of an additional mechanism. The concept of alertness has been central in the scholarly efforts to uncover the mechanisms underlying opportunity discovery. Nevertheless, the fuzziness of the concept has also been acknowledged. For example, Alvarez and Barney described it as ‘a subject that has long eluded entrepreneurship scholars’ (Alvarez and Barney, 2002, p. 93), a situation that is arguably precluding the field from advancing more quickly and more significantly. The concept of alertness was originally suggested by Kirzner (e.g. 1973, 1979, 1997, 1999, 2009). Kirzner himself recently recognized that ‘certain significant misunderstandings’ about his conceptualization of alertness subsist in the literature (2009, p. 148). In our view, Kirzner seems to equate ‘alert individuals’ with ‘entrepreneurs’ (Shane, 2003, p. 10). In consequence, while many studies in this domain tried to capture a single view on alertness, it seems that he was actually referring to different types of mechanisms for opportunity formation across his significant body of research. Our paper helps to clarify some of these misunderstandings.

First, one conceptualization of alertness refers to discovery without search (Kirzner, 1979). This kind of discovery seems possible through identification: the focal exogenous shock generates enough information to enable actors to estimate the expected value of the opportunity (Shane and Venkataraman, 2000). But what happens if the information available is insufficient for the expected value computation? What mechanisms can reduce the level of ignorance to allow some decision-making? Our perspective suggests that in such cases automatic interpretation should be the expected mechanism, and the effects on decision-making and action should be those predicted in our entrepreneurial interpretation model.

Second, one possible mechanism for discovered opportunities would be directed search. Due to his early intention to position his arguments apart from Stiglerian ideas, Kirzner's body of work and subsequent Kirznerian-based studies tended to ignore directed search (and stressed undirected search instead). Kirzner argued that actors would not know ‘what to look for’ (1997, p. 72). In fact, he was right: directed search requires that some other mechanisms operate before in order to guide the subsequent search. Our model clearly shows such a mechanism: opportunity interpretation. Opportunity interpretation (as depicted in our Figure 1) allows the determination of which information is missing. Actors would know ‘what to look for’. Moreover, our model points out specific conditions under which directed search would be expected: that is, when the level of interpretive resolution is moderate, and ex post ignorance is perceived as mainly caused by unknown missing information.

Third, by bringing directed search back to opportunity discovery, our paper also clarifies the boundaries of undirected search. Following Kirzner's early suggestions, several streams of research tended to emphasize (albeit in varying tones) undirected search (e.g. Fiet, 2007; Kaish and Gilad, 1991; Ozgen and Baron, 2007), which was another conceptualization of ‘alertness’. Some studies suggested a more focused kind of search than those usually considered. Fiet (2007), for example, proposed a constrained search through deliberate choices (and grouping) of sources of information. It is unclear whether this kind of search belongs to directed or undirected search. Our arguments suggest that such a mechanism falls under the category of undirected search since it assumes focal actors know ‘where to look’ rather than ‘what to look for’. In fact, it is consistent with another (and less commonly used) definition of alertness, in this case viewed as the ‘propensity to know where to look for information’ (Kirzner, 1973, p. 68; emphasis added).

Implications for the Creation View

In contrast to the discovery perspective, creation is a relatively recent theoretical view. Not surprisingly, important issues concerning this view remain unaddressed. Our study illuminates important aspects of some of those issues. First, we propose the inclusion of a new mechanism, constructed interpretation. Research to date has tended to ascribe creation processes to situations of complete ignorance. This is likely to be the case at the beginning of the creation process since it involves creating something from nothing (Baker and Nelson, 2005). However, as the process develops, significant information is likely to be produced, and as a result the nature of the context will also evolve. Specifically, the informational setting is likely to be subject to a transition from complete ignorance to a situation characterized by both equivocal and incomplete information. In such a context, interpretation is likely to be required. In fact, ‘perceptions’ and ‘beliefs’ have been already mentioned in past studies (Alvarez and Barney, 2007; Sarasvathy, 2001).

Little work, however, has been undertaken so far to include cognitive representations of available information in creation processes. We address this gap by suggesting a specific mechanism, constructed interpretation. We show that, as happens to any interpretation in general, this mechanism enables the transformation of equivocal and missing information into an understanding of the environment that in turn allows subsequent decision making and actions. But, we also argue, this particular mode of interpretation contrasts with automatic interpretation since it involves engaged actors that participate in a continuous and effortful process, rather than actors that behave as external observers in a one-shot and essentially mindless process. In addition, consistent with the social nature generally ascribed to creation process, we expect that it is the collective (rather than the individual) interpretation that will be consequential for subsequent actions.

Second, we provide detailed reasoning that portrays creation as a moderately structured process (where collective interpretations play a crucial role). Prior research within the creation tradition argues that orderly sequential processes might be rare in entrepreneurship (Baker and Nelson, 2005). While we fully acknowledge the intensively iterative nature of a creation endeavour, our theory suggests that some stages might be discernible in the process. Specifically, whereas previous works emphasized action and reaction (e.g. Alvarez and Barney, 2007), we propose that action, individual interpretation, and shared understandings (along with specific modes of reaction) are at work in successive steps. Initial (usually blind) action generates information that, given its equivocal and incomplete nature, requires interpretation by individual actors. Yet, as a social constructed process, individual interpretations lead to shared understandings or collective interpretations, which represent consensus or agreement (in varying degrees) among the involved actors. A decision to exploit is expected only when the collective interpretive resolution is high, i.e. when actors are able to show a high degree of agreement that the situation fits quite well the actors' opportunity templates. A low collective interpretive resolution (that is, a high degree of agreement that the situation does not fit the actors' opportunity template) is likely to lead to an abandonment decision.

When the collective interpretive resolution is moderate, our model predicts directed action, directed search, or interpretive absorption actions depending on the perceived dominant cause of ex post ignorance (the ignorance prevailing even after the activation of the interpretive mechanism). These predictions demonstrate two additional contributions made by this study. On the one hand, we suggest that directional action might occur in creation processes more often than usually considered. Adner and Levinthal (2008) recently proposed that observers tend to evaluate some entrepreneurial actions as having a random nature (while they are indeed rather directed) because those actions pursue dimensions of performance not already legitimated in a given organizational context. Our model extends their suggestion by adding a complementary reason for such ‘misunderstanding’ about directed actions: some directed actions are guided by prior activation of focal actors' collective interpretation, which depend on the actor-specific opportunity templates, and on the specific dynamics underlying the development of such shared understandings; as different groups of people may have different collective knowledge structures, observers' ‘misunderstanding’ may be due to discrepancies in knowledge structures. Moreover, whereas Adner and Levinthal (2008) acknowledged that actors' ‘hunches and hypotheses’ might serve as triggers of directed action, our model specifies that it is the confluence of moderate levels of interpretation resolution and the predominance of unknowable ex post ignorance that are more likely to originate directed action.

On the other hand, our model also predicts two distinct modes of ‘reaction’. Research has stressed the importance of reaction in addition to action. Reaction would mean that actors try to obtain from their strategic peers a response to the outputs of their actions (Sarasvathy, 2001). We expand this reasoning by showing that two different modes of reaction should be expected. One, directed search, means that actors search for specific information that they perceive as missing, and should occur when the interpretive resolution is moderate and ex post ignorance is dominated by unknown missing information. Another mode of reaction, interpretive absorption actions, means that actors, rather than looking for information, attempt to share the knowledge structures held by third parties that could enrich their reinterpretation of the situation. This kind of effort is expected when the interpretive resolution is moderate and ex post ignorance is dominated by equivocality.

CONCLUSION

In sum, this paper introduces an overlooked but crucial mechanism for opportunity formation (entrepreneurial interpretation), and develops a novel theoretical model of its effects on a set of outcomes such as decision to exploit, interpretive absorption actions, directed search, directed action, and inaction. We further elaborate on the use of entrepreneurial interpretation to address several important questions that remained unanswered in discovery and creation views, the two dominant theoretical perspectives on opportunity formation. We hope this paper helps to provide some crucial missing pieces in the fascinating puzzle known as the entrepreneurship theory.

ACKNOWLEDGMENTS

I thank Andrew Corbett and the three anonymous reviewers for their insightful feedback and encouragement. I gratefully acknowledge David Patient, Marta Magalhães, Diogo Barreto, and JMS editorial staff for their valuable help.

Ancillary