Skip to main content Accessibility help
Hostname: page-component-5c569c448b-ph4cd Total loading time: 0.514 Render date: 2022-07-02T08:40:09.223Z Has data issue: true Feature Flags: { "shouldUseShareProductTool": true, "shouldUseHypothesis": true, "isUnsiloEnabled": true, "useRatesEcommerce": false, "useNewApi": true } hasContentIssue true

Using analogies to explain versus inspire concepts

Published online by Cambridge University Press:  27 April 2015

Amanda Chou
Department of Mechanical and Industrial Engineering, University of Toronto, Toronto, Canada
L.H. Shu*
Department of Mechanical and Industrial Engineering, University of Toronto, Toronto, Canada
Reprint requests to: L.H. Shu, Department of Mechanical and Industrial Engineering, University of Toronto, 5 King's College Road, Toronto, ON M5S 3G8, Canada. E-mail:
Rights & Permissions[Opens in a new window]


We aim to examine the potential of using analogies in design education and to compare the roles of analogies in explaining versus inspiring in engineering design. We review existing research in analogical thinking, with a focus on scientific discourse and education. Then we explore the role of analogies in design education in making concepts more relatable by asking six participants in a graduate-level design course to generate analogies for course topics. We describe criteria developed to evaluate the analogies and present these evaluations. We then asked participants to perform divergent thinking tests, but we found no significant correlation between these and analogy scores. The participants were also asked to reflect on what constitutes an effective analog, describe their process of identifying analogies, and provide their definitions of analogies. We describe possible links between these comments and the ratings of their analogies. We then draw on results in using analogies in pedagogy to inform and reflect on obstacles we encountered in the use of analogies to inspire. Specifically, we related them to our experience with biomimetic or biologically inspired design, where we used a natural-language search approach to identify relevant analogies. Three aspects discussed are familiarity of source analogies, boundaries of parallels between source analogies and target concepts, and concreteness of source analogies. Finally, we discuss possible pedagogical benefits of eliciting analogies on course topics from students, namely, using the elicited analogies as tools for improved student engagement as well as more prompt instructor feedback.

Special Issue Articles
Copyright © Cambridge University Press 2015 


Analogies can be used to explain a foreign target concept through a more familiar source (or analog, anchor, base, and vehicle) concept by identifying similarities in structure and relationships. For design, source analogies are used to inspire target concepts. More generally, analogies represent a cognitive process whereby information from one concept is transferred to another. While our laboratory's experience has been in the use of analogies to inspire, we became interested in their use to explain (i.e., clarify concepts through detailed descriptions) and found that insights gained in the latter may inform our observations in the former. We begin with an overview of the study of analogies that focus on their role in explaining.

Analogical reasoning has played an important role in science, law, and even politics. For example, Harré (Reference Harré and Hilton1988) stated that many analogies were used in explanations in medicine and physics. Holyoak and Thagard (Reference Holyoak and Thagard1995, Reference Holyoak and Thagard1997) provide examples of analogy use in fields from psychotherapy to politics. Duit (Reference Duit1991), Clement (Reference Clement1993), and others reveal the potential in analogies to familiarize abstract concepts by grounding them in real-world examples, provide visualizations, and even motivate students by connecting a foreign concept with a familiar, perhaps fun, idea. Duit and Treagust (Reference Duit and Treagust2003) also note that analogies can lead to conceptual change learning, whereby the learner experiences a restructuring of preinstructional mental models, or understanding, to learn a new concept. Analogies have also played important roles in significant scientific advancements; Nersessian (Reference Nersessian2008) established that model-based reasoning, such as creating analogies, forms the basis for novel representations of concepts that push the development of science.

1.1. Clarification of analogical expressions

Below are brief definitions of analogical expressions that often appear in educational contexts and everyday speech. By themselves, these are informal vehicles for delivering analogical thought, and their use highlights pitfalls with careless use of analogies to explain.

1.1.1. Metaphor versus simile

Both metaphors and similes are considered analogical figures of speech. These terms are used more commonly in the arts than to describe formal mappings between scientific concepts. In the classroom, however, analogies are often phrased as these rhetorical figures of speech. Generally, metaphors assert that one thing is another, while similes highlight their similarity using “like” or “as.” For example, “imagine that electricity is water flowing through a conduit,” is a metaphor, while “electricity is like water flowing through a conduit” is a simile.

1.1.2. Idioms

An idiom is a combination of words with a figurative meaning that cannot be determined from the sum of the individual meanings of its constituents. For example, “spill the beans” means to let out a secret, which has nothing to do with spilling or beans. Many idioms are analogical, and mappings between the figurative meaning and literal statement can be constructed; for example, “beans” are like secrets, and when spilled, are difficult to clean up, like how a secret cannot be untold.

Unlike explicit formulation of analogies to explain a scientific concept, idioms are used as colloquialisms that the speaker or author expects the audience to know, much like any other word on its own. When uttering “spill the beans,” the speaker is unlikely to explain the reference to secrets. Thus, the use of idioms assumes the audience already knows the mapping, or at least the intended figurative meaning. Such an assumption may be inappropriate in education, because explaining a foreign concept using an analog that the student does not understand makes little sense. Furthermore, both language skills and personal tendencies have significant effects on the appropriateness of idioms as effective communication; for example, the many engineers who study and work in a nonnative language, and are more literal in nature, may be particularly ill suited for idioms.

1.2. Analogies and cognition

At a cognitive level, some believe that analogies are an inherent way of thinking. Lakoff and Johnson (Reference Lakoff and Johnson1980) argue that metaphors are a fundamental mechanism of the mind; concepts that govern our thought also govern our everyday functions, and the human conceptual system is largely metaphorical. Gentner (Reference Gentner, Vosniadou and Ortony1989) suggests that in the processing of sensory information from external stimuli, the brain searches through analogical “bundles” for comprehension. In other words, what we know about something depends on everything else we know. Heywood (Reference Heywood2002) also suggests that learning via analogies creates a deeper appreciation for knowledge, as it is rooted in socially constructed meaning as opposed to absolute “truth.”

We note that analogies are a natural way to explain concepts. For example, when asked, “What is ___?” we often say, “it is like . . . ,” and perhaps add “but . . .” or “however, . . . .” We root our understanding of things on other things, learn by building connections, and explain using connections between ideas.

1.3. Analogies as hypotheses

When two things share parallels in many aspects, there are likely to be further parallels (Gentner, Reference Gentner2002). However, while analogies can help elucidate a new concept, if similarities are extrapolated too far, understanding of the target can be compromised, or entirely wrong. In contrast, because the mapping between similar aspects of the concepts is symmetrical, the target concept often introduces a new perspective for the source concept. That is, we can learn more about the source through what we know about the target, and not just the other way around. Treagust et al. (Reference Treagust, Duit, Joslin and Lindauer1992) refer to this as the “two-way aspect” of analogies.

Many great figures in science and engineering attest to the value of analogical thinking in the discussion of new concepts. They conceptualize, communicate, and even advance their ideas and discoveries through analogies. Analogies can not only communicate and explain concepts but also identify structural similarities between domains and further scientific discovery by helping in forming hypotheses. In the discovery of a novel concept A, and its mapping to a similar analog concept B, we can postulate that perhaps A shares more aspects of B beyond what we already know.

Podolefsky and Finkelstein (Reference Podolefsky and Finkelstein2006) note that James Clerk Maxwell (Reference Maxwell and Niven1885), a mathematical physicist best known for his equations on electromagnetism, applied mathematical ideas obtained from fluid mechanics to electrical science, and used analogies to generate physical hypotheses. Davies et al. (Reference Davies, Nersessian and Goel2005) present a detailed analysis of visual analogies used by Maxwell. By drawing potential additional parallels, Reynolds et al. (Reference Reynolds, Pease and Li2004) constructed an algorithm to generate hypotheses based on structural mapping of analogies, leading to plausible hypotheses given sets of aligned structural statements.

Scientific advancement often comes from the investigation of these hypotheses. Theoretical physicist and Nobel laureate Richard Feynman argues that analogies enable physicists to maintain a broad knowledge despite numerous ongoing discoveries; many equations share analogous mathematical forms, indicating similarities between physical phenomena (Feynman et al., Reference Feynman, Leighton and Sands1964). Stephen Hawking uses 74 analogies in A Brief History of Time (Reference Hawking1988). Other analogies are Carnot's heat engines with waterfalls and Rutherford's planetary model of the atom (Karam & Ricardo, Reference Karam and Ricardo2011).

1.4. Overview of paper

The remainder of this paper is organized as follows. Section 2 reviews the use of analogies in science education. Section 3 presents an exploratory study on the role of analogies in design education. Section 4 draws on results in using analogies in pedagogy to inform and reflect on obstacles we encountered in the use of analogies to inspire. Section 5 identifies potential pedagogical benefits of eliciting analogies on course topics from students.


In science, similes and metaphors help form the mental models necessary for students to understand new concepts (Solomon, Reference Solomon1986). Explanations often require references to nonobservable entities, which often rely on analogies (Harré, Reference Harré and Hilton1988).

2.1. Classroom use of analogies

Several researchers studied the use of analogies in the classroom. Clement (Reference Clement1993) analyzed student understanding of concepts taught using analogies, found a lack of care in choosing analogies to teach physical concepts, and suggested that a series of bridging analogies, each with smaller conceptual gaps, aids in explaining distant concepts. Clement also recommends interactive group discussions, as students often contribute creative examples and can evaluate whether the presented analogies make sense to them. Treagust et al. (Reference Treagust, Duit, Joslin and Lindauer1992) studied how teachers incorporate analogies into their class preparation and found, in a sample of 7, that high school teachers used examples more than analogies, and some were unable to differentiate between the two. Analogies, when used, were often not used to their full potential (i.e., lacking in detail).

2.2. Textbook use of analogies

Noting that over half of the analogies found in textbooks are not explained, and many that are, are done with little depth, Duit (Reference Duit1991) suggests that guidance on analogies in textbooks is also needed, but often missing. Orgill and Bodner (Reference Orgill and Bodner2006) note that concrete analogies tend to be used to illustrate abstract concepts in biochemistry textbooks, for example, “the triple helix structure is similar to that of a rope” (Boyer, Reference Boyer1999). Concreteness can also refer to intangible, but familiar concepts. For example, “in the energy economy of a cell, glucose reserves are like ready cash” (Campbell, Reference Campbell1999) assumes a concrete grasp of the availability of cash versus other financial resources.

2.3. Limitations of analogies in education

Acknowledging that analogies break down once the conceptual similarities are exhausted, Gentner and Gentner (Reference Gentner, Gentner, Gentner and Stevens1983) note that certain analogies may be more relevant in explaining particular components of a new concept. For example, they found that to understand electric circuits, students using a “moving crowd” analogy understood the concept of resistors better, while those using a “flowing water” analogy understood the concept of batteries better. Thus, self-formed analogies, or mental models of how things work, can both inform and misguide students. If cognition is a bundle of analogies, the network of existing bundles will determine how new pockets of information will be incorporated within the students’ larger mental models. Clement et al. (Reference Clement, Brown and Zietsman1989) found in a study that examples that experts deemed to be highly analogous to a given phenomenon were often not understood well by students. Students may have unasked questions, and make their own assumptions, extrapolating from the analogy. Harrison and Treagust (Reference Harrison and Treagust1996) found that high school science students often transferred attributes from the analog literally, and that analogical models can be dangerous when students are left to draw their own conclusions from them.

2.4. Analogical models for education

After Zeitoun's (Reference Zeitoun1984) general model of analogy teaching, many others established their own frameworks, including Glynn's (Reference Glynn, Glnn and Britton1991) teaching with analogies model, designed to help guide teachers in using analogies in a systematic way. Both Zeitoun and Glynn recognized that care must be taken in the use of analogies in the classroom, and that limitations must be defined. Because the concepts are inherently different, Glynn confirms that all analogies break down somewhere, as the mapping is never exactly one-to-one. Thus, to avoid misinterpretation, nonshared attributes must be discussed. Glynn also recognizes that analogies should be chosen to accommodate different backgrounds to be familiar to many students, as analogies only work when rooted in existing knowledge (Glynn & Muth, Reference Glynn and Muth1994).

2.5. Analogies in design versus science education

The study of analogies in science education appears focused on physical phenomena, where a nonobservable mechanism is explained using an analogous, observable mechanism. For example, to explain static normal force, Clement (Reference Clement1993) begins with a hand on a vertical spring to explain that a table can exert a normal force on a book resting on top. While it is easy to see that a spring can push back on an applied force, it is less obvious that a table can do the same.

Design methods relate more to humans than to fundamental laws, but analogies can improve accessibility of a concept, method, or guideline by relating it to ideas familiar to students. Analogies can also reveal higher order relations that allow students to extend concepts beyond the domain of design. Because many concepts taught in design education are mental tools and guides, we believe that by providing a broader analogical context for a course concept, design tools can be more effectively and readily understood, recalled, and used by students.


In light of the above insights, we wanted to explore the role of analogies in design education by conducting a series of classroom experiments. The participants were six master's students (two female, four male) in a graduate-level course on creativity in conceptual design offered by the Department of Mechanical and Industrial Engineering at the University of Toronto. Of the six students, five had backgrounds in mechanical engineering, and one had a background in civil and environmental engineering. Two were working full time while pursuing course-based master's degrees part time, two were pursing research-based master's degrees, and the remaining two were pursuing course-based degrees as full-time students.

3.1. Eliciting cross-domain analogies for course concepts

We asked students to write down and submit cross-domain analogies for the main concept taught, at the end of each lecture. Most students generated only one analogy per concept, and generally took 5–10 min. After noting that many of the first set of analogies addressed limited aspects of a concept, we asked participants to identify analogies that support the benefit of a method, approach, or guideline, for example, the benefit of generating several concepts early in the design process. Later, we also asked for analogies that reveal the limitations of concepts, for example, lead-user methods. These analogies were then discussed in the subsequent lecture. Appendix A shows examples of “good” analogies for each of 10 course topics.

3.2. Rating analogies

We developed a rubric of three criteria, each with four levels, shown in Table 1. Logic addresses whether the structural relationships in the proposed analogy match the concept. It also includes factuality, as we noticed that some statements about biological analogies were inaccurate. Clarity addresses the quality of communication. Because we received the responses in written form, vagueness and overly brief or convoluted statements compromised raters’ understanding. Finally, domain addresses whether the analogy reveals a different perspective, or is a literal example. For example, to support the benefit of modular architecture, one participant wrote about site trailers that could be connected together, which was deemed a same-domain example.

Table 1. Rubric for scoring analogies generated on course concepts

Our criteria for assessing analogies for design course concepts differ from those suggested by Gentner (Reference Gentner and Miall1982) for scientific analogies. Detailed structural mapping is emphasized in explanatory analogies for scientific concepts, but we emphasize the ease of seeing a connection between the domains. This is closer to what Gentner (Reference Gentner and Miall1982) refers to as expressive analogies.

3.3. Results and examples of rating analogies

Two raters independently scored all 118 analogies generated for 10 course concepts. We computed the Cohen κ, a statistical measure of interrater agreement for categorical ratings, and achieved a “moderate” level of agreement (Landis & Koch, Reference Landis and Koch1977) for all three criteria: κlogic = 0.53, κclarity = 0.47, and κdomain = 0.45. We found responses had to perform well (score 3–4) in all three criteria to correspond to an intuitively good analogy; for example, a clearly explained response with poor logic and distinct domain is still ineffective. Examples follow.

  1. 1. Example of a 444-rated analogy (for need identification methods):

    River flow measurementuse of various methodsstanding in river with flowmeter and taking care to stand far enough away so as not to interfere flow, but could be some interference still; launching a “free” flowing device and measuring/receiving telemetry from outside the river.

  2. 2. Example of a 141-rated analogy (for modular architecture) clear, but not an analogy:

    If one part breaks down, just replace that one part instead of whole thing, e.g., car parts, tennis strings vs. getting new racquet

  3. 3. Example of 323-rated analogy (for functional decomposition), unclear mapping of function/needs:

    A family who wants to move to a new location. First, they have to identify what the functions (needs) of the family members such as education, recreation, work. And decomposing the functions (needs) to specify the details such as what type of school? To help the family decide what locations have the best functions & characteristics for the family to move into.

Table 2 provides each participant's scores for each criterion averaged over all the participant's analogies, the standard deviations, and the average score for the three criteria.

Table 2. Mean analogy scores and standard deviations for all six participants

We noticed that participants were fairly consistent in their exhibited approach. Analogies that were rated higher, where the analogical components were clearly outlined and structural mappings were logical, tended to consistently come from the same participants. Other participants tended to generate vague analogies, capturing the concepts only minimally. We note that Participant 3 scored well consistently, while Participant 4 performed poorly, but produced 5.8 analogies per concept on average. All other participants typically produced one or two responses, with Participant 6 occasionally producing three responses per concept/topic.

3.4. Discussion of elicited analogies

In generating analogies for “affordances,” two students described medicinal side effects to express that affordances of a product may not have been intended by the designer. Participant 3's analogy below may help students appreciate the potential severity of consequences arising from disregarding affordances in design. However, the analogy also clarifies the limitations of affordances; while side effects can take over the main purpose of the drug, as is the case with ViagraTM, affordances cannot overtake the main functions of a product.

Affordances are like … side-effects when designing pharmaceutical drugs. E.g., Bayer researchers designed aspirin and heroin to fulfill the function of painkilling but the unintended side-effects of heroin led to it becoming an addictive recreational drug. Therefore identifying affordances is important in preventing unintentional harm to users (or leading to unethical benefits for the company by getting users addicted).

Highly rated analogies were often presented in a structured manner, with analogous components mapped explicitly. For example, in the last sentence of the below, Participant 3 maps form to function, suggesting that explicit mapping could create better analogies in analogy generation.

The benefit of identifying/decomposing functions in a football analogy would be the coach organizing and planning strategies with the attackers and defenders on the team to achieve those objectives. For example, the positioning of the defenders would be the “form” to achieve the “function” of preventing opponents scoring.

In participant comments on what makes analogies effective (Appendix B), students mention personally relatable ideas, as well as low effort in understanding. This corresponds well with our evaluation criteria, which addresses ease of understanding. Gick and Holyoak (Reference Gick and Holyoak1983) noted that analogies not familiar to students become an extra burden to learn and fail to explain or describe.

3.5. Tests of divergent thinking

We next conducted creativity tests to see whether their results could be correlated with quality of analogies generated. While logical reasoning (often measured as IQ) is required for analogical correctness, generating analogies is also a divergent thinking task involving creativity. The domains from which analogies can be drawn are theoretically unlimited. Thus, we postulate that those with higher quality analogies may score higher on divergent thinking tasks.

Guilford (Reference Guilford1967) suggests the following measures for creativity: fluency (number of responses); flexibility (number of response categories); originality (number of unique responses); and elaboration (level of detail of responses).

We sought to examine whether creativity scores between tasks involving simple, everyday ideas align with creativity scores in analogical generation, and whether these scores could indicate high-quality analogies. Because we elicited the analogies with minimal prompt (only asked participants to generate analogies for a given topic), we can apply these same measures to the responses. However, simpler, universally familiar creativity exercises, for example, generate drawings using circles as prompts, may have fewer constraints. Therefore, the reduction in number or range of ideas in analogy generation tasks could be due to fewer viable ideas.

To avoid bias toward particular problem domains and favoring students familiar with those domains, the Guilford tests were performed using open-ended questions with loose restrictions. The Miller Analogies Test and Remote Associates Test (Mednick, Reference Mednick1963) were also considered, but they were deemed overly dependent on participant vocabulary size.

Tasks asked of participants are as follows, where parenthesized information was not provided as text, but describes the worksheets that were provided.

  • Use the (grid of four columns by five rows of) circles as prompts for drawing. Each circle is used in a separate drawing. Draw for 5 min.

  • Use the (grid of four columns by five rows of) triangles as prompts for drawing. Each triangle is used in a separate drawing. Draw for 5 min.

  • Name as many uses for a binder clip as you can think of in 5 min. Use bullet points.

  • Name as many uses for a candle as you can think of in 5 min. Use bullet points.

3.6. Results

Below, we discuss the results of comparing creativity measures and analogy generation.

3.6.1. Fluency

Fluency, or number of responses, is the most objective of Guilford's suggested creativity measures. We compare average fluency values for analogy generation, alternative uses, and shape-prompted drawings. Spearman's rank correlations for each pair are shown in Table 3.

Table 3. Spearman's correlations for fluency between analogy generation and creativity tasks

Table 3 shows that only fluency between listing alternative uses and drawing are highly correlated and statistically significant. This suggests that with unrestrictive creative activities, the relative volume of ideas produced by participants is more consistent, regardless of whether the task is verbal or visual. With analogy generation, the extra constraints and logical analyses required of the task may reduce the number of analogies produced.

3.6.2. Originality and flexibility in alternative uses test

We quantify originality only for the alternative uses test. Generated analogies were almost always drawn from different domains, as were shape-prompted drawings, which were not restrictive enough to cause many overlaps in ideas between participants. Alternative uses, in contrast, had several common as well as unique responses. Two raters identified all unique responses for all participants, with uniqueness defined as inversely proportional to frequency of occurrence in the sample. Interrater agreement was measured using the Cohen κ (κ = 0.44), which, according to Landis and Koch (Reference Landis and Koch1977), is a qualitative “moderate agreement.”

The same two raters measured flexibility for the alternative uses test. Each rater developed a set of categories for the entire set of responses, and enumerated the categories appearing in each participant's answer. The Pearson correlation coefficient of the flexibility scores given by each rater is ρ = 0.95. The results of the drawing exercise are discussed anecdotally below.

3.6.3. Analogy generation versus creativity tests

We next examined our data by comparing analogy generation to creativity test results. With six participants, we only suggest qualitative trends.

Alternative-uses tasks

Alternative uses often had varying specificity, which resulted in overlaps in listed uses. For example, for uses of candles, one participant listed “burn,” “warmth,” and “melt plastic,” so there is at least one overlap as “burn” refers to either burning of the candle itself (warmth) or use of the candle to burn another object (melt plastic). We did not take this into account in the scoring; while overlaps may be accounted for in low elaboration scores (because detail in responses will reduce likelihood of overlap), alternative uses results were of similar elaboration levels. We noted that all participants produced more uses for a candle than a binder clip, perhaps due to its ease of phase change, and having more common everyday uses (e.g., birthday cakes are mentioned by all but one participant).

Drawing tasks

For the drawing questions, there were also no distinct differences between degrees of elaboration, except possibly by Participant 5. Elaboration and fluency appear to come at the cost of originality and flexibility, and ideas often follow the same theme, such as sports, food, or nature. There were no obvious differences in fluency between the problem types, and the number of responses may indicate the approach participants take in tackling problems.

Participant 3

Participant 3 consistently produced high-quality analogies in class, and scored second highest in fluency for alternative uses, though notably low in flexibility. In the drawing questions, although not following the provided instructions, this participant demonstrated creative use by incorporating multiple circles in two of the drawings. Participant 6 was the only other one who used more than one shape per drawing, though in a much more abstract manner. Participant 3 is also noted for high flexibility in analogy generation, covering topics such as river flow, pharmaceutics, and football. However, it remains unclear whether the ability to generate spontaneous ideas given few restrictions correlates with generating good analogies.

Participant 4

Of note, Participant 4 consistently produced the most responses for the drawing tasks. However, this participant's responses were also consistently unelaborated, and exhibited low flexibility in both analogy generation and drawing questions, gathering ideas from a limited range of topics. The participant admitted later to the belief that all shapes must be filled for the drawing questions, though this was not indicated in the problem statement. In the alternative user test, Participant 4 was flexible and obtained the highest creativity quotient score (Snyder et al., Reference Snyder, Mitchell, Bossomaier and Pallier2004). This is consistent with Kudrowitz and Dippo (Reference Kudrowitz and Dippo2013), who identified that those with more responses tended to also have more novel responses. Participant 4 also mentioned after the test that a previous course on innovative product design provided exposure to insightful uses of a candle.

Kudrowitz and Dippo (Reference Kudrowitz and Dippo2013) noted that novel ideas tended to arise after the first nine responses in the alternative uses test. However, we did not see this in our limited sample size, even though our participants were given 5 min for each alternative uses test, versus 3 min in their study.

3.6.4. Analogy generation versus definitions of analogy and processes to identify analogies

Finally, we asked participants to provide definitions of analogies as well as describe the process they underwent when identifying analogies. Appendix B contains excerpts of responses. In hindsight, we believe that these responses may explain the difference in the ratings of analogies. For example, the lowest-rated analogies were generated by those who characterize the process as “random” (Participants 4 and 6 average scores = 2.29 and 2.53). Next lowest-rated analogies came from those who emphasized the importance of “emotion,” that is, emotional compellingness, in an analogy (Participants 5 and 2 average scores = 2.80 and 3.03). Highest scores were 3.42 and 3.10 (Participants 3 and 1). All scores are out of 4.00.


By examining studies on the use of analogies in education, we identify insights that may explain obstacles in the application of analogies in biomimetic design. We refer to our experience with biological analogies retrieved through our natural-language search tool (Shu, Reference Shu2010), developed to aid engineers in generating design solutions. We compare the use of analogies in the classroom for purposes of explaining a new concept with the mapping of biological phenomena to some engineering problem and solution domain.

4.1. The analogy must be understandable (if not familiar)

Many works attempting to identify teaching strategies to ensure effective use of analogies, for example, the Teaching-With-Analogies model, suggest that the analog concept used to explain should be a familiar one. Dagher and Cossman (Reference Dagher and Cossman1992) refer only to analogies as the part of explanatory discourse where the familiar is used to explain the unfamiliar.

In biomimetic design, the biological analogies that inspire the most are likely to be unfamiliar to the designer, yet we informally observed that both novice and expert designers are biased toward familiar analogies. In practice, we believe that designers would be quite familiar with the problem with which they are tasked before they start looking for sources of inspiration in biology or elsewhere. However, given likely limitations in the designer's knowledge of analogous phenomena, the bias toward familiarity, in both source and target domains, may limit opportunity for inspiration. Specifically, we have long argued that designers should look beyond the scales or organization levels in biology with which they are most familiar, that is, from organ (e.g., heart, lung, or hand) to organism (e.g., specific animal or plant), and can most readily be observed casually (Shu et al., Reference Shu, Ueda, Chiu and Cheong2011). In addition, we observed that even expert designers tend to match novel analogies to existing and known solutions, rather than develop new solutions. We note that in our past experiments (Cheong & Shu, Reference Cheong and Shu2013b), where novice designers were asked to generate design concepts given an engineering problem and a biological phenomenon, both concepts were presented as new information. Thus, mapping was perhaps more difficult in these experiments as opposed to a setting where the designer can seek alternative sources to better understand the concepts prior to mapping. In addition, novice designers tended to fixate on specific words within the stimulus text, particularly familiar words, for example, “motor” of “motor protein,” and develop nonanalogous solutions, for example, those that incorporate motors (Shu et al., Reference Shu, Ueda, Chiu and Cheong2011).

Therefore, the question remains: how do we nudge designers away from the natural bias for the familiar, and toward analogies in the form of unfamiliar phenomena that may serve to inspire more? Similarly, how do we nudge designers toward developing new solutions from unfamiliar analogies, rather than mapping them to known solutions?

Researchers in biologically inspired design are working toward answering the first question by recognizing the difficulty engineers may have in understanding biological analogies and making the analogies as clear as possible, through pictures, diagrams, and so forth (Goel et al., Reference Goel, McAdams and Stone2014).

4.2. Limits of analogy should be defined

Models for systematizing analogy use in classrooms also suggest that the limits of the analogical boundary should be made clear to the student. Specific to written analogies, Glynn and Takahashi (Reference Glynn and Takahashi1998) note that explicit mapping of the analog and target concepts must be made, in addition to explicit statements of limits of the analogy and conclusions that can be drawn in the target domain. In our classroom experiment, we separately sought analogies that support a concept, as well as demonstrate limitations of the concept. Experiment participants found value in this process, as well as in discussing the merits and limitations of example analogies they generated. Two different concepts cannot have a complete one-to-one mapping: inevitably, the analogy breaks down beyond the shared parallels. In other words, there must be extraneous aspects of each concept that do not map to the other. This can lead to students extrapolating parallels where they do not exist, which can misguide their understanding of the target concept.

Again, this is problematic when designers are provided a biological analogy for use to solve an engineering problem. In a previous study (Cheong et al., Reference Cheong, Hallihan and Shu2014), senior undergraduate engineering students were asked to design a credit card marketing solution using the pheromone release of ants as an analogy. Some students continued to draw parallels between the analogy and problem by stating that the queen ant was analogous to the CEO of the credit card company. This is an example of an unnecessary parallel because it does not contribute to the design solution, and the queen ant was not mentioned in the description of the biological phenomenon provided.

We note, however, analogical mappings to inspire are inherently open-ended, as designers must extrapolate from the biological phenomenon to their design solution by first identifying parallels between the domains. Because they may draw from multiple source analogies, extraneous information adds to the complexity of the task. In other words, there may be aspects of the biological phenomenon that have no appropriate parallel in the design problem, but will be tempting to draw. Goel (Reference Goel, Vattam, Wiltgen and Helms2012) notes that rerepresentation (e.g., rephrasing of text) of the domains is often necessary for an analogy to become obvious. This further adds to the difficulty in drawing analogies for design inspiration. Helms et al. (Reference Helms, Vattam and Goel2009) also found that novice designers tend to fixate on biological features not analogous to the problem.

Designers may benefit from further researching the retrieved biological phenomenon in order to better understand its mechanisms prior to mapping. That is, designers may need to actively identify where the analogies break down to avoid fixating on irrelevant attributes of the phenomenon. In our laboratory, Cheong and Shu (Reference Cheong and Shu2013b) developed templates to support the mapping of relevant attributes of source and target domains. Perhaps support, for example, guidelines, in identifying irrelevant attributes, could also be developed and provided in addition to those to identify relevant attributes.

4.3. Concrete examples work best

Concrete source analogies seem to be favored in explaining complicated ideas. Lemke (Reference Lemke1990) suggests that science teachers attempt to humanize science by associating it with social processes in order to aid student understanding. Biochemistry textbooks mainly use concrete analogies to inform abstract concepts (Orgill & Bodner, Reference Orgill and Bodner2006), for example, glove and hand for enzyme and substrate. After analyzing several hundred science textbooks and comparing the relative levels of abstraction between the source and target concepts, Orgill and Bodner noted that analogies tend to be used in cases where the target material is difficult or abstract. While most target concepts are abstract, most analogs used to explain are concrete in nature, which are easier to understand (Curtis & Reigeluth, Reference Curtis and Reigeluth1984; Thiele & Treagust, Reference Thiele and Treagust1994).

The preference for concreteness over abstraction is supported by our observation of many challenges encountered by novice designers when asked to map a biological phenomenon to an engineering problem. Cheong and Shu (Reference Cheong and Shu2013b) substituted biological nouns with more abstract nouns, for example, “blood” with “fluid,” in text descriptions to reduce fixation and bias when selecting and mapping biological analogies. However, this intervention did not improve the quality of mappings between the biological phenomena and the engineering solutions. We now see that in eliminating the possibility of designers using previous knowledge of biological concepts, we also eliminate any possible familiarity with the notions. The concepts become more difficult for designers to both understand and map, which was confirmed to statistical significance (Feng et al., Reference Feng, Cheong and Shu2014). In essence, abstraction forces the designer to use only the remaining information available to create analogies. While this may be a helpful exercise for analogical thinking, it is generally unheard of in educational contexts.

However, Gentner and Jeziorski (Reference Gentner, Jeziorski, Gholson, Shadish, Neimeyer and Houts1989) identify analogies as relational commonalities between entities, which are independent of the attributes of the objects holding the relation. In the solar system versus atom example, we note that the hotness of the sun is irrelevant to the relationships between the analogous components. In this sense, abstracting nouns (e.g., sun) to prevent designers from incorporating irrelevant information (e.g., hotness) makes sense. Table 4 summarizes insights from use of analogy to explain, to the use of analogy to inspire.

Table 4. Comparison between analogy use in pedagogy and as design inspiration


A study of analogies in education provided us with another perspective on the challenges we observed in designers when using biological phenomena as design stimuli. In Table 4, we identify three recurring points in use of analogies for both explaining and inspiring, particularly for novices: use concrete analog concepts that are easy to picture, identify boundaries at which the analogy breaks down, and use understandable if not familiar analog concepts.

We note that designers may have to learn more about the retrieved biological phenomenon to first gain a better grasp of the mechanism itself, before trying to map it to the design problem. The designer must become familiar with both the design problem and the stimulus. That is, rather than attempt to reduce fixation on familiar analogies, or familiar words in the description of biological phenomena (as we have attempted), we should instead ensure that all relevant analogies are as understandable as possible, which is no minor feat. We have not been able to implement automated solutions that outperform Google and Google Images to aid in understanding of biological phenomena. However, these extra steps may be obstacle enough to direct designers to the most familiar, rather than the most useful or novel, analogies.

In our small study of analogy generation and divergent thinking in a graduate-level design class, we noticed that some students consistently produced high-quality, logical analogies, while others consistently produced lower quality analogies. We sought to examine the relationship between divergent thinking tasks and analogy generation by administering drawing exercises and alternative uses tests. We found a high, significant correlation between fluencies in drawing and alternative uses listing, but neither correlated with analogy generation. In hindsight, we believe that how a participant defines analogies and approaches the process of generating analogies may have a larger effect on the quality of his or her analogies. Longitudinal studies with a larger sample may allow us to identify and assess ways to improve analogical generation and reasoning.

5.1. Pedagogical benefits of eliciting analogies on course topics

While we did not obtain statistically significant results in our exploratory study of analogies in teaching, we did notice some unexpected benefits of eliciting analogies on course topics.

5.1.1. Analogy as feedback for instructor

Due to the small sample size, we did not formally compare the participants’ class scores with their analogy scores. However, we did find that Participants 3 and 4 tied for the highest class scores but performed, respectively, the best and the worst in analogy generation. Prince and Felder (Reference Prince and Felder2006) note that traditional engineering methods are deductive, where examples are used to illustrate general concepts, but inductive teaching is at least as, if not more, effective. Thus, we postulate that engineering students such as our participants may be accustomed to deductive teaching and evaluation methods, which may result in course grades being poor predictors for analogy generation skills.

However, poor analogies generated by all participants could indicate an unclear lecture. For example, analogies elicited for fixation revealed a common confusion between fixation and other cognitive biases, and led to an additional unscheduled lecture on cognitive biases. The elicitation of analogies for a lecture topic to be discussed in the following lecture may provide feedback to the instructor more promptly than more formal assessment exercises.

5.1.2. Analogy as engagement tool

Analogies can serve as engagement tools by linking concepts to ideas or activities students find interesting or enjoyable. For example, we previously observed positive student responses to sports analogies, for example, “design fixation is like skating with one's head down in hockey,” and “not learning multiple methods for the same task is like playing tennis with only a forehand.” Because interest differs between students, inclusiveness becomes an issue. For example, the use of sports analogies likely works better with athletic or sports-minded students. Nonetheless, analogies catering to student interests may function beyond explanation, serving to inspire interest in topics pertaining to science and engineering. One benefit of eliciting analogies from students, which are subsequently discussed in class, is that such analogies are personal to the students and may enable inclusion that is difficult to achieve with preselected analogies. For example, an analogy for lead-user methods suggesting that Ramadan (a period of fasting) can serve to inform how dieters can lose weight was quickly rebutted by others familiar with the religious practice.

In an end-of-term debriefing session, the participants identified additional benefits of identifying analogies for course topics. Multiple students expected that they would find the course concepts taught to be more memorable because they would recall their own as well as their peers’ analogies. Several students found that the process of identifying analogies became easier over time, and they reported their experience with analogies in teaching to be positive. Therefore, we will investigate how to incorporate the use of analogies in other courses with more participants.


The authors thank the participants of the study and the Natural Sciences and Engineering Research Council of Canada. This study was approved by the University of Toronto Ethics Review Board (Protocol Reference No. 30303).


Examples of “good” analogies for design course topics


Selected excerpts from comments on analogy generation and analogy definitions

Amanda Chou is an MASc candidate in mechanical and industrial engineering at the University of Toronto, where she completed her undergraduate degree in engineering science: infrastructure. Her interests are natural language processing and its applications in product design.

L.H. Shu is a Professor of mechanical engineering at the University of Toronto. She received her SM and PhD degrees in mechanical engineering from the Massachusetts Institute of Technology. Dr. Shu is a Fellow of the CIRP (International Academy of Production Engineering) and has taken leadership roles in the Design Society and the American Society of Mechanical Engineers Design Theory and Methodology Committee.


Boyer, R. (1999). Concepts in Biochemistry. Pacific Grove, CA: Brooks/Cole.Google Scholar
Campbell, M.K. (1999). Biochemistry, 3rd ed.Orlando, FL: Harcourt Brace.Google ScholarPubMed
Cheong, H., Hallihan, G.M., & Shu, L.H. (2014). Design problem solving with biological analogies: a verbal protocol study. Artificial Intelligence for Engineering Design, Analysis and Manufacturing 28(1), 2747.CrossRefGoogle Scholar
Cheong, H., & Shu, L.H. (2013 a). Reducing cognitive bias in biomimetic design by abstracting nouns. CIRP Annals 62(1), 111114.CrossRefGoogle Scholar
Cheong, H., & Shu, L.H. (2013 b). Using templates and mapping strategies to support analogical transfer in biomimetic design. Design Studies 34(6), 706728.CrossRefGoogle Scholar
Clement, J. (1993). Using bridging analogies and anchoring intuitions to deal with students' preconceptions in physics. Journal of Research in Science Teaching 30(10), 12411257.CrossRefGoogle Scholar
Clement, J., Brown, D., & Zietsman, A. (1989). Not all preconceptions are misconceptions: finding anchoring conceptions for grounding instruction on students' intuitions. International Journal of Science Education 11, 554565.CrossRefGoogle Scholar
Curtis, R., & Reigeluth, C. (1984). The use of analogies in written text. Instructional Science 13, 99117.CrossRefGoogle Scholar
Dagher, Z., & Cossman, G. (1992). Verbal explanations given by science teachers: their nature and implications. Journal of Research in Science Teaching 29(4), 361374.CrossRefGoogle Scholar
Davies, J., Nersessian, N., & Goel, A. (2005). Visual models in analogical problem solving. Foundations of Science 10(1), 133152.CrossRefGoogle Scholar
Duit, R. (1991). The role of analogies and metaphors in learning science. Science Education 75(6), 649672.CrossRefGoogle Scholar
Duit, R., & Treagust, D. (2003). Conceptual change: a powerful framework for improving science teaching and learning. International Journal of Science Education 25(6), 671688.CrossRefGoogle Scholar
Feng, T., Cheong, H., & Shu, L.H. (2014). Effects of abstraction on selecting relevant biological phenomena for biomimetic design. Journal of Mechanical Design 136(11), 111111.CrossRefGoogle Scholar
Feynman, R.P., Leighton, R.B., & Sands, M. (1964). The Feynman Lectures on Physics, Vol. 2. Reading, MA: Addison–Wesley.Google Scholar
Gentner, D. (1982). Are scientific analogies metaphors? In Metaphor: Problems and Perspectives (Miall, D.S., Ed.), pp. 106132. Brighton: Harvester Press.Google Scholar
Gentner, D. (1989). The mechanisms of analogical learning. In Similarity and Analogical Reasoning (Vosniadou, S., & Ortony, A., Eds.), pp. 199241. London: Cambridge University Press.CrossRefGoogle Scholar
Gentner, D. (2002). Psychology of Analogical Reasoning. Encyclopedia of Cognitive Science. London: Nature Publishing Group.Google Scholar
Gentner, D., & Gentner, D.R. (1983). Flowing waters or teeming crowds: mental models of electricity. In Mental Models (Gentner, D., & Stevens, A., Eds.). Mahwah, NJ: Erlbaum.Google Scholar
Gentner, D., & Jeziorski, M. (1989). Historical shifts in the use of analogy in science. In Psychology of Science: Contributions to Metascience (Gholson, B.W., Shadish, R., Neimeyer, R., & Houts, A., Eds.), pp. 296325. New York: Cambridge University Press.CrossRefGoogle Scholar
Gick, M.L., & Holyoak, K.J. (1983). Schema induction and analogical transfer. Cognitive Psychology 15(1), 138.CrossRefGoogle Scholar
Glynn, S.M. (1991). Explaining science concepts: a teaching-with-analogies model. In The Psychology of Learning Science (Glnn, S.R., & Britton, B., Eds.), pp. 219240. Hillsdale, NJ: Erlbaum.Google Scholar
Glynn, S.M., & Muth, K.D. (1994). Reading and writing to learn science: achieving scientific literacy. Journal of Research in Science Teaching 31(9), 10571073.CrossRefGoogle Scholar
Glynn, S.M., & Takahashi, T. (1998). Learning from analogy-enhanced science text. Journal of Research in Science Teaching 35(10), 11291149.3.0.CO;2-2>CrossRefGoogle Scholar
Goel, A.K., McAdams, D.A., & Stone, R.B. (Eds.). (2014). Biologically Inspired Design: Computational Methods and Tools. London: Springer–Verlag.CrossRefGoogle Scholar
Goel, A.K., Vattam, S., Wiltgen, B., & Helms, M. (2012). Cognitive, collaborative, conceptual and creative—our characteristics of the next generation of knowledge-based CAD systems: a study in biologically inspired design. Computer-Aided Design 44(10), 879.CrossRefGoogle Scholar
Guilford, J.P. (1967). The Nature of Human Intelligence. New York: McGraw–Hill.Google Scholar
Harré, R. (1988). Modes of explanation. In Contemporary Science and Natural Explanation: Common-Sense Conceptions of Causality (Hilton, D.J., Ed.). New York: New York University Press.Google Scholar
Harrison, A., & Treagust, D. (1996). Secondary students' mental models of atoms and molecules: implications for teaching chemistry. Science Education 80(5), 509534.3.0.CO;2-F>CrossRefGoogle Scholar
Hawking, S. (1988). A Brief History of Time. New York: Bantam Books.Google Scholar
Helms, M., Vattam, S., & Goel, A. (2009). Biologically inspired design: process and products. Design Studies 30(5), 606622.CrossRefGoogle Scholar
Heywood, D. (2002). The place of analogies in science education. Cambridge Journal of Education 32(2), 233247.CrossRefGoogle Scholar
Holyoak, K.J., & Thagard, P.R. (1995). Mental Leaps: Analogy in Creative Thought. Cambridge, MA: MIT Press.Google Scholar
Holyoak, K. J. & Thagard, P. (1997). The analogical mind. American Psychologist 52(1), 3544.CrossRefGoogle ScholarPubMed
Karam, R., & Ricardo, E.C. (2011). Formal analogies in physics teacher education. Proc. Int. ESERA Conf. 2011, Lyon, France.Google Scholar
Kudrowitz, B., & Dippo, C. (2013). Getting to novel ideas: exploring the alternative uses test of divergent thinking. Proc. ASME IDETC/CIE, Paper No. DETC2013-13262, Portland, OR.Google Scholar
Lakoff, G., & Johnson, M. (1980). Metaphors We Live By. Chicago: University of Chicago Press.Google Scholar
Landis, J.R., & Koch, G.G. (1977). The measurement of observer agreement for categorical data. Biometrics 33(1), 159174.CrossRefGoogle ScholarPubMed
Lemke, J. (1990). Talking Science. Norwood, NJ: Ablex.Google Scholar
Maxwell, J.C. (1885). Abstract of paper “On Faraday's Lines of Force.” In The Scientific Papers of James Clerk Maxwell (Niven, W.D., Ed.), pp. 367369. Cambridge: Cambridge University Press.Google Scholar
Mednick, M.T. (1963). Research creativity in psychology graduate students. Journal of Consulting Psychology 27(3), 265266.CrossRefGoogle ScholarPubMed
Nersessian, N.J. (2008). Creating Scientific Concepts. Cambridge, MA: MIT Press.Google Scholar
Orgill, M., & Bodner, G.M. (2006). An analysis of the effectiveness of analogy use in college-level biochemistry textbooks. Journal of Research in Science Teaching 43(10), 10401060.CrossRefGoogle Scholar
Podolefsky, N.S. & Finkelstein, N.D. (2006). The use of analogy in learning physics: the role of representations. Physical Review ST: Physics Education Research 2(2), 020101.Google Scholar
Prince, M.J., & Felder, R.M. (2006). Inductive teaching and learning methods: descriptions, comparisons, and research bases. Journal of Engineering Education 95(2), 123138.CrossRefGoogle Scholar
Reynolds, J., Pease, A., & Li, J. (2004). Analogy and deduction for knowledge discovery. Proc. IEEE 2004 Int. Conf. Information and Knowledge Engineering, Las Vegas, NV.Google Scholar
Shu, L.H. (2010). A natural-language approach to biomimetic design. Artificial Intelligence for Engineering, Design Analysis and Manufacturing 24(4), 507519.CrossRefGoogle Scholar
Shu, L.H., Ueda, K., Chiu, I., & Cheong, H. (2011). Biologically inspired design. CIRP Annals—Manufacturing Technology 60(2), 673693.CrossRefGoogle Scholar
Snyder, A., Mitchell, K., Bossomaier, T., & Pallier, G. (2004). The creativity quotient: an objective scoring of ideational fluency. Creativity Research Journal 16(4), 415420.CrossRefGoogle Scholar
Solomon, J. (1986). Children's explanations. Oxford Review of Education 12(1), 4151.CrossRefGoogle Scholar
Thiele, R.B., & Treagust, D.F. (1994). An interpretive examination of high school chemistry teachers' analogical explanations. Journal of Research in Science Teaching 31(3), 227242.CrossRefGoogle Scholar
Treagust, D.F., Duit, R., Joslin, P., & Lindauer, I. (1992). Science teachers' use of analogies: observations from classroom practice. International Journal of Science Education 14(4), 412422.CrossRefGoogle Scholar
Zeitoun, H.H. (1984). Teaching scientific analogies: a proposed model. Research in Science and Technological Education 2(2), 107205.CrossRefGoogle Scholar
Figure 0

Table 1. Rubric for scoring analogies generated on course concepts

Figure 1

Table 2. Mean analogy scores and standard deviations for all six participants

Figure 2

Table 3. Spearman's correlations for fluency between analogy generation and creativity tasks

Figure 3

Table 4. Comparison between analogy use in pedagogy and as design inspiration

You have Access
Cited by

Save article to Kindle

To save this article to your Kindle, first ensure is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about saving to your Kindle.

Note you can select to save to either the or variations. ‘’ emails are free but can only be saved to your device when it is connected to wi-fi. ‘’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

Find out more about the Kindle Personal Document Service.

Using analogies to explain versus inspire concepts
Available formats

Save article to Dropbox

To save this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you used this feature, you will be asked to authorise Cambridge Core to connect with your Dropbox account. Find out more about saving content to Dropbox.

Using analogies to explain versus inspire concepts
Available formats

Save article to Google Drive

To save this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you used this feature, you will be asked to authorise Cambridge Core to connect with your Google Drive account. Find out more about saving content to Google Drive.

Using analogies to explain versus inspire concepts
Available formats

Reply to: Submit a response

Please enter your response.

Your details

Please enter a valid email address.

Conflicting interests

Do you have any conflicting interests? *