Entrepreneurial Cognition

by Dean A Shepherd

  To illuminate the difference between superficial features and structural relationships, my (Dean) colleagues and I (Grégoire et al. 2010: 416) used an example of a new technology which had been developed at MIT—namely, the 3DP™ discussed in Shane (2000). We illustrated the differences between superficial features and structural relationships as follows:Examples of superficial features of the technology include who developed the technology (mechanical engineers at MIT), the components of the technology (mechanical arm, print head), the material it uses (ceramic powders), and what the technology produced in the lab (e.g., ceramic filters, casting molds, etc.). Examples of first-order structural relationships include how the technology operates (e.g., [mechanical arm (moves) print head]; [print head (deposits) powder]). Higher-order structural relationships include more abstract capabilities of the technology (e.g., [how the technology operates] causes [fabrication of tridimensional objects with high level of automation and precision]).

  Structural-alignment processes are a significant aspect of individuals’ sensemaking efforts regarding new information. When individuals are presented with a new stimulus, they evaluate how its features and relationships align with those of a pertinent “source” (Gentner 1989; Holland et al. 1986). For example, this source could be a related object, or it could be a more intangible framework including a category or theoretical model the individual holds. Yet, more frequently such a related object is a mental representation of a situation that informs the individual’s understanding of the new information. This comparison of new information with a related model, object, or situation (i.e., the determination of whether a target’s superficial features and structural relationships align with those of a source) enables individuals to detect patterns that convey meaning. Based on these patterns, the individuals can derive useful conclusions.

  However, research on structural alignment has also shown that different sets of cognitive structures and dynamics are required to process superficial features and structural relationships (Gentner 1989; Keane et al. 1994). Consequently, these two aspects of structural alignment are likely to impact opportunity-identification efforts in different ways. On the one hand, superficial features affect how individuals search for and retrieve information from their memory (Gentner 1989; Gentner et al. 1993). Thus, a new stimulus’s superficial features (e.g., the material required for operating a new technology) may trigger individuals’ recollection of comparable features of an important source (e.g., a market offering the material referenced in the previous example). The source the individual recalls from memory is often shaped by his or her previous experiences or familiarity with specific features. Alternatively, it can be shaped by his or her environment or current situation (e.g., a feature may be salient for a person because of particular events in his or her life). This shaping limits the number of superficially related domains one instantly (and unconsciously) accesses (Keane et al. 1994) when scanning for pertinent references for alignment. Structural relationships, in contrast, are intertwined in a more direct manner with higher-order processes of reasoning (Keane et al. 1994). As an example, the processing and alignment associated with structural relationships affect individuals’ formation of categories (Namy and Gentner 2002), solving of problems (Catrambone and Holyoak 1989, 1990), and learning (Loewenstein and Gentner 2005).

  Superficial features and structural relationships can both impact people’s interpretations. However, scholars have demonstrated that structural relationships are especially crucial when individuals make inferences about a stimulus that is novel and/or ambiguous (Day and Gentner 2007; Gentner 1989). As such, my (Dean) colleagues and I (Grégoire et al. 2010) theorized that people’s attempts to identify opportunities are likely to stress their use and alignment of structural relationships. There are two notions underlying this stress. First, individuals are likely to draw on structural relationships when stimuli are encoded in a deep and rich way. For example, such deep and rich encoding happens when one performs a cognitively demanding or emotionally challenging task (Blanchette and Dunbar 2001; Catrambone and Holyoak 1989). The opportunity-identification task fulfills both these conditions as information required is typically ambiguous and difficult to interpret. Similarly, these tasks are usually emotionally engaging largely due to the possible outcomes they may yield for entrepreneurs and their firms (Cardon et al. 2012; Ireland et al. 2003).

  Second, scholars have found that from a neuro-cognitive perspective, the brain is activated more when individuals perceive alignment of structural relationships than when they notice alignment of superficial features (Holland et al. 1986; Keane et al. 1994). Based on this partiality for the alignment of structural relationships, individuals are better able to identify and compare meaningful patterns. These patterns might include superficial similarities (but not necessarily). Indeed, researchers from several fields have documented individuals’ use of structurally based “mental leaps” (Holyoak and Thagard 1996). Such mental leaps occur, for example, when individuals think creatively and/or attempt to solve scientific problems (e.g., Dahl and Moreau 2002; Dunbar 1993; Ward 1995). In the context of strategic decisions, Gavetti and Rivkin (2005) reported how Andrew Grove (the former Intel CEO) realized the risk of deserting the low-end microprocessor segment. Instead of considering the context of computer or electronic products, he related Intel’s situation to what occurred in the steel sector after Nucor and mini-mills were introduced. While reinforcing bars and microprocessors do not share many comparable characteristics, Nucor’s entrance and success in the steel business was very similar to Intel’s entrance and success in the microprocessor sector. Thus, Grove was able to formulate a strategy that prevented Intel from experiencing a similar future because he had knowledge of the history and decline of established US steel companies. The discussion above implies that when trying to identify opportunities, individuals tend to pay more attention to the alignment of structural relationships than to the alignment of superficial features. Along these lines, my (Dean) co-authors and I revealed that the opportunity-identification process requires higher levels of cognitive effort (i.e., attention) for the alignment of structural relationships than for the alignment of superficial features (Grégoire et al. 2010).

  The Role of Prior Knowledge in the Structural-Alignment Process

  As mentioned earlier , scholars have shown that since knowledge is not evenly distributed throughout the population, prior knowledge provides at least a partial explanation as to why some people are able to identify specific opportunities that other people miss (e.g., Corbett 2005; Dimov 2007b; Shane 2000; Shepherd and DeTienne 2005). Overall, work in this area has argued that prior knowledge serves as a foundation for the interpretation and use of new information; however, most studies on this topic have not delineated the cognitive mechanisms by which prior knowledge affects individuals’ opportunity recognition . We believe that prior knowledge likely triggers individuals’ consideration of structural relationships. For instance, domain experts often find reasoning in terms of structural relationships easier because they can draw on deeper mental representations (Chi et al. 1981). Such experts are particularly good at solving problems characterized by low levels of superficial similarity but high levels of structural similarity (Keane 1988). Additionally, research has demonstrated that when people fail to solve particular problems, “failure indices” are frequently left in long-term memory. Following Seifert et al.’s (1994) “opportunistic-assimilation hypothesis,” these indices remain inactive until one has an encounter with a stimulus related to addressing the problem. At that point, failure indices “serve as signposts that guide subsequent retrieval processes back to stored aspects of the problematic situation” (Seifert et al. 1994: 87). That is, prior experience with a problem can enhance an individual’s attentiveness to stimuli that are relevant for finding a solution (Dimov 2004). What these perspectives demonstrate is that prior knowledge enables individuals to notice structural similarities between new information and relevant contexts although sup
erficial connections between the two are missing.

  In line with the reasoning above, my (Dean) colleagues and I (Grégoire et al. 2010) revealed that in the opportunity-recognition process, individuals’ dependence on higher levels of prior knowledge requires more cognitive effort (i.e., attention) for the alignment of structural relationships than for the alignment of superficial features. The results uncovered were in line with a structural-alignment model of opportunity identification, suggesting that these cognitive processes are vital to identifying opportunities. My (Dean) colleagues and I (Grégoire et al. 2010) showed that when entrepreneurs came across information related to a novel technology, they focused on the parallels between this information and contexts in which it could be useful. Further, the structural-alignment process involves various types of similarities, each having different outcomes, and some of these similarity considerations encompass the superficial features of technologies and markets. In line with studies in cognitive psychology (e.g., Gentner 1989; Keane et al. 1994), the findings imply that these features direct individuals’ early attempts to look for contexts that serve as a point of reference for assessing the significance of the stimulus (which in our example includes the identification of markets that may align with the technology). Yet, most attempts to make sense of new information and determine whether the technology and the market fit in a way that they constitute a possible opportunity primarily depend on considering and aligning structural relationships (Grégoire et al. 2010).

  Most importantly, the study found that perceiving similarities between higher-order relationships seems to be a vital part of the opportunity-identification process, a notion that is supported by three additional lines of evidence. First, when participants made verbalizations highlighting similarities between the superficial features of technologies and superficial features of markets, they allocated a significantly higher amount of attention to the alignment of the structural relationships between technologies and markets, and in doing so they emphasized high-order structural relationships. Moreover, several times the entrepreneurs came up with opportunities when there were numerous structural relationships that were shared by technologies and markets but when technologies and markets had few common superficial features. Said differently, entrepreneurs’ ability to notice the alignment of structural relationships enabled the transfer of technologies across domains and thus formed opportunity beliefs that were not overtly evident. Third, when the entrepreneurs placed more emphasis on a stimulus’ superficial features rather than on its structural relationships, they had greater difficultly thinking of possible opportunities. Similar challenges surfaced when other matters inhibited the entrepreneurs’ ability to consider structural relationships (e.g., when one participant concentrated on assessing the viability of obtaining intellectual property protection for the technology or when another participant focused on time limits). As a whole, the three lines of evidence highlight that while superficial features may direct individuals’ initial thinking about new information, reasoning based on the aligning structural relationships is a vital part of opportunity recognition.

  Differences in the Nature of Opportunities and the Structural-Alignment Process

  The discussion above concentrated on the factors that enable some individuals or organizations to identify and act upon promising activities (cf. Gruber et al. 2010; Plambeck and Weber 2009; Short et al. 2010). Although there has been sustained interest in and theorizing about the nature and sources of opportunities (e.g., Alvarez and Barney 2010; Jackson and Dutton 1988; McMullen et al. 2007), scholars have paid less theoretical and empirical attention to the impact of differences across opportunities, particularly in regard to initial opportunity identification. However, my (Dean) colleague and I (Grégoire and Shepherd 2012) created and tested an opportunity-identification model focusing on the effects of differences across potential opportunities. Expanding the assumptions outlined above, opportunity beliefs form as a result of cognitive efforts to understand possible “matches” between new ways of supply (e.g., new services, products, technologies, or business models) and the markets in which these new means of supply can be introduced. Thus, in the context of technology transfer, the formation of opportunity beliefs hinges on the consideration entrepreneurs give to the structural alignment between new technologies and markets (as described above and specified in Grégoire et al. [2010]).

  The Effects of Convergent and Divergent Variations in Alignment

  When thinking about structural alignment, we need to take into consideration that superficial and structural similarities can differ independently of one another. From a modeling perspective, the question thus arises as to whether the effects of superficial and structural similarity are merely additive or whether these two forms of similarity interact with one another. To answer this question, my (Dean) colleague and I (Grégoire and Shepherd 2012) tested for a possible interaction between the two dimensions (as detailed below). However, when trying to understand the challenges associated with recognizing potential opportunities, it becomes especially important to explore the meaning and influence that differences across forms of alignment may have on the development of opportunity beliefs. This issue is particularly relevant when the superficial and structural similarities of a technology-market combination are at odds with each other.

  While new technologies are often depicted as only being appropriate for specific applications (i.e., how the technology was utilized “in the lab”), entrepreneurs frequently envision other uses for technologies in entirely different markets than the inventors (or those in charge of the commercialization) originally had in mind. Indeed, Shane (2000) described how the opportunities envisioned for the technology he investigated were frequently “non-obvious” even to entrepreneurs trying to exploit other opportunities for the same technology. Explanations for this “non-obviousness” have generally emphasized the role of entrepreneurs’ unique knowledge resources in this context. More specifically, due to their greater knowledge and understanding of particular markets and industries compared to many technology inventors, some entrepreneurs are able to recognize market applications that the inventors never could have imagined (Gruber et al. 2008, 2012; Shane 2000; Ucbasaran et al. 2009).

  Over and above entrepreneurs’ prior market knowledge, a complementary explanation for opportunity identification is focused attention on the distinct influence of superficial and structural similarity in the development of opportunity beliefs. In the context of our model (Grégoire and Shepherd 2012), the seeming non-obviousness of opportunities appears to be caused by divergences stemming from the low levels of superficial similarities shared between markets and technologies even though they share high levels of structural similarities.

  Cognitive science researchers have found that the human mind prefers reasoning involving higher-order structural relationships when interpreting ambiguous stimuli in uncertain contexts (Gentner 1989; Holland et al. 1986). For example, when making predictions about new objects, people generally prefer predictions that proceed from a comprehensive causal system as opposed to predictions that—while equally conceivable—are not part of such a system (Clement and Gentner 1991). Similarly, studies have shown that structural matches usually lead to more brain activity compared to superficial matches because the former activate more neuronal connections (Keane et al. 1994). The implication here is that when individuals think about entrepreneurial opportunities, they are likely to be more cognitively “aroused” when they notice commonalities between a new technology’s structural features and the causes of latent demand in a market than when they notice superficial similarities between the technology and the market.

  Despite the mind’s preference for structural similarities, recognizing and processing structural similarities without superficial parallels are especially demanding cognitive activities (cf. Catrambone and Holyoak 1989). As a result, the absence of superficial similarities characteristic of some technology-market combinations can make opportunity ideas l
ess apparent even when a technology’s capabilities correspond to the causes of latent demand in a market. In turn, individuals may feel less certain or less positive about the resulting opportunity beliefs than they would in the case of high superficial similarity. Students often experience this challenge, for example, they often have a difficult time transferring the content and solutions they learn in one domain with specific superficial elements (e.g., math problems that use particular objects or units) to other domains with logically similar problems but different superficial features (e.g., physics problems focusing on different objects and units) (cf. Bassok and Holyoak 1989; Novick and Holyoak 1991). Ultimately, the absence of superficial similarities often makes knowledge transfer more challenging.

  On the other hand, a dominant focus on superficial similarities can at times result in flawed reasoning premises, such as when there are superficial similarities present without structural similarities. For example, strong similarities between a technology’s superficial elements and a market could potentially offset the detrimental effects of structural discrepancies between the technology’s capabilities and the causes of latent demand in the market. When this occurs, individuals’ opportunity beliefs are likely to be less negative than they would have been without such strong superficial similarities.

  As a whole, these observations could explain why casual observers find it difficult to identify opportunities comprising technologies that share low levels of superficial similarity but high levels of structural similarity with markets. Again, while the human brain favors making inferences based on structural relationships, identifying and processing such relationships when superficial parallels are lacking are cognitively demanding. Nevertheless, cognitive researchers have shown that low superficial/high structural reasoning is vital to making inferences that enhance knowledge in highly uncertain contexts (Holland et al. 1986) and to making creative “mental leaps” (cf. Holyoak and Thagard 1996), such as when scientists, engineers, designers, and strategists come up with imaginative solutions to complex problems (Dahl and Moreau 2002; Dunbar 1993; Gavetti and Rivkin 2005).