Analogies and Analogical Reasoning in Invention

Affiliation
Ulf Krumnack Kai-Uwe Kühnberger Angela Schwering
University of Osnabrück University of Osnabrück University of Münster
Osnabrück Osnabrück Münster

Tarek R. Besold
University of Osnabrück
Osnabrück

Synonyms

Invention, Creativity, Analogy

(I) The Concepts “Analogy” and “Invention”

Analogy
An analogy is usually considered as a structural mapping between a source (base) and a target domain. To
establish an analogy, common substructures of the two domains are identified and mapped to each other,
resulting in an analogical relation. The establishment of an analogy is usually governed by certain
constraints, like systematicity, structural consistency, or a one-to-one restriction on possible mappings,
although there is no general accepted set of such principles. Analogy-making can also be regarded as the
establishment of a generalization, identifying an abstract core that consists of the common structures of
both domains and ignores surface appearance and domain peculiarities (cf. Fig. 1). Analogies are usually
not judged right or wrong, rather they can be more or less plausible, based on the degree of structural
coherence that they exhibit, possibly depending on the context and analogical conclusion they allow to
draw.

Fig. 1: Analogy via Generalization

Analogical reasoning uses an analogical relation to transfer structures, relational properties, or

conceptions from the source to the target domain. Parts of the source domain that have no correspondence
on the target side, are candidates for transfer and can be fitted into the target by translating them based on
the analogical mapping. Unlike other inference mechanisms that basically make implicit knowledge about
a domain explicit, analogical reasoning can import new conceptions into the target domain. It is important
to notice that an analogical inference does not result in factual knowledge, even when source and target
domain consist of definite facts. It is the task of the reasoner, to carefully examine the results, may it be
for logical consistency, or for empirical evidence, before accepting them as part of knowledge about the
target domain.
Analogies have been identified as a core mechanism of human cognition. The recall of memories based
on current impressions, the understanding of a new situation in terms of a familiar one, the creation of
abstract concepts, or the ability to learn from quite restricted sets of examples are most likely based on
analogy-making. Especially, analogical reasoning allows for the introduction of new ideas into a domain
and thereby provides an explanation of human creativity.
A wide variety of models for analogical thinking have been introduced, which differ in the way they
represent knowledge the analogy is based on and in the methods used to establish the analogical relation
and to realize the analogical transfer. On the symbolic side, the Structure Mapping Theory (SMT) uses
graphs to represent the domains and computes analogical relations by identifying maximal matching
subgraphs. The logic-based Heuristic-Driven Theory Projection (HDTP) represents domain theories by
sets of axioms and generalizes matching subtheories. The copycat model of Hofstadter's fluid analogy
group is based on the notion of conceptual slippage, and analogies are found by adjusting the
representation of source and target. Kokinov's DUAL/AMBR architecture realizes its memory as a
network of micro-agents, in which analogies are found by a hybrid process of spreading activation and
marker passing. There are also models which are based on connectionist and subsymbolic representations.
Holyoak and Hummel created a framework for Learning and Inference with Schemas and Analogies
(LISA) that is based on a neural network model that uses temporally synchronized activation between
neurons to signify a mapping between source and target elements. Vector Symbolic Architectures (VSAs)
use high-dimensional vectors which can be accessed by simple mathematical operations, to store, retrieve
or analogically reason about information.

Invention
Creativity, innovation, and invention are three abstract concepts that are strongly interrelated and can be
used for specifying the process of how to generate new ideas, products, or solutions to problems. Whereas
creativity describes a general cognitive capacity that is in different degrees involved in any process of
generating an invention, the concepts invention and innovation describe properties of concrete products,
services, or ideas. From a more engineering and business oriented perspective, an invention is usually
considered as the manifestation of the creative mental act, resulting in a new artifact (prototype), a new
type of service, a new concept, or even the mental concretization of a conception. An innovation requires
standardly the acceptance of the invention by the market, where market is not exclusively restricted to
business aspects. Therefore, there can be non-inventive innovations (e.g. a product that has market
success, but is not pioneering in any sense), non-innovative inventions (e.g. a pioneering product that has
no market success), and also inventions that lead finally to innovations (e.g. a pioneering product that has
market success) (Burki and Cavalluci 2011).
Inventions are also important milestones for the progression in art, culture, and science. Inventions in
such fields are not necessarily coupled with concrete products or services, rather insights of the artists or
scientist cause the development of new concepts and new ideas.
 Examples for scientific inventions are the discovery of the complex plane in mathematics, enabling a
geometric interpretation of algebraically defined complex numbers (Argand 1813), Einstein’s theories
of special and general relativity, or Chomsky’s insight that natural language can be described by a
recursive mechanism (phrase structure grammar) (Chomsky 1957). Although, the term discovery
(instead of invention) is most often used in the science context, it is not undisputed whether scientific
research discovers existing entities (e.g. Platonic ideas) or just invents models that approximate and
explain reality. In this article the term invention is also used for the scientific domain.
 Examples for artistic inventions are numerous. Not only the invention of new forms (types) of art,
like, in the case of music, the form of a symphony starting in the 17th century or rock and roll music in
the mid-20th century, but also many other types of inventions like the building of new instruments in
music, new techniques for painting, new methods to compose texts (sometimes together with other
forms of media content, like pictures) etc. can be subsumed under invention.
Creativity as the underlying cognitive ability to produce inventions in science and art is, similar to the
cases mentioned above, the necessary cognitive prerequisite for pioneering developments. Nevertheless,
the term innovation, although obviously usable and important also in science and art, is only indirectly
measurable contrary to many business products and services.

(II) The Role of Analogy in Invention

Invention in Engineering and Business Processes

An explicitly dedicated framework for a theory of invention in the engineering and business domain is
TRIZ, the Russian acronym (“теория решения изобретательских задач”) for the Theory of Inventive
Problem Solving. TRIZ was developed in its original form by Genrich Altshuller, a Russian inventor who
examined more than 40,000 patent abstracts in order to find common patterns and principles behind
business relevant inventions (Altshuller 1984). TRIZ is based on the idea that at the beginning of an
inventive problem solving process a technical or physical contradiction arises. According to TRIZ this
contradiction can be resolved by applying 40 principles of invention (mostly parameter settings in
engineering problems) relative to an ideal system. The general architecture behind the solution process is
depicted in Fig. 2.

Fig. 2: The general TRIZ architecture of an inventive problem solving process

An interesting feature of TRIZ is the explicit lifting of a given problem (contradiction) to an abstract
level, in order to apply the principles of inventive problem solving. The lifting process itself can be
understood as an analogy-driven process, because similar problems need to be consulted to find the right
generic problem category (compare Fig. 1). In this sense, TRIZ uses analogy-making in a non-trivial
sense for generating inventive solutions.
Besides TRIZ, many problem set-ups for innovation processes, invention creation, and the design of
inventive products are strongly related to optimization problems. A framework originally developed for
the statistical assessment of optimality in quality management is 6. In order to make already existing
processes adequate for optimization, the so-called DMAIC cycle proposes five steps: define, measure,
analyze, improve, and control such processes. For each step 6 proposes tools for the optimization
process. Nowadays, 6is widely used in manufacturing industry and the financial sector.
Also computational methods for inventive problem solving have been proposed. As an example case-
based reasoning (CBR) can be mentioned, originating from research in artificial intelligence. The idea of
CBR is to build a new solution for a given problem (domain) by using solutions for similar problems that
are stored in a knowledge base. The algorithmic process is divided into four steps: (i) retrieve and (ii)
reuse a similar solution stored in the knowledge base, (iii) refine it to a new solution, and finally (iv)
retain the new solution. The appropriateness of the new solution is strongly dependent on the chosen
similarity measure between the problems. Quite often CBR is mentioned directly as an instance of
analogical reasoning. Although there are significant differences between the analogy concept specified
above and CBR (e.g. the lack of the systematicity principle in CBR or lack of a transfer process of
knowledge from source to target), it is rather clear that both methods are related to each other and CBR
can be interpreted as strongly constrained form of analogy-making.

Invention in Science and Mathematics

Throughout history, analogies have been reported as driving force for invention and scientific discovery.
Holyoak and Thagard (1995) mention that already in ancient times the Roman engineer Vitruvius
observed similarities between water waves and sound. Later in the seventeenth century an analogy
between sound and light was established, giving rise to the wave theory of light. Also Rutherford's atom
model, based on an analogy to the solar system, although physically disproven, is still used in education,
because it is easy to conceptualize and appropriate for a first conception of a new domain.
Analogies are also seen indispensable in the formation of mathematical concepts. Georg Polya (1954)
emphasizes the central role of analogies in the development of mathematical concepts and proofs. Based
on linguistic evidence, Lakoff and Núñez (2000) describe the development of mathematical ideas as a
long series of analogies (“conceptual metaphors” in the terminology of the authors), starting with
grounding metaphors which relate basic arithmetic operations to physical actions in the world. From these
basic notions further linking metaphors give rise to ever more abstract concepts like infinity, transfinite
numbers, and the continuum via generalization and blending. In a case study, Martinez et al. (2012)
describe Argand's invention of complex numbers interpreted as vectors in the complex plane as a network
of analogical mappings and conceptual blend spaces between arithmetic and geometric ideas. Similarly to
the business processes and engineering problems, Argand started with a contradiction: There was no
possibility to give complex numbers an appropriate interpretation on the (real) number line.
In engineering, new technical devices have often been introduced based on analogies to biological and
physical systems. Bell designed his telephone based on an analogy with the human ear. The Wright
brothers got inspiration from soaring birds when they developed aircraft and de Mestral invented the
velcro fastener by imitating burdock burrs. The whole discipline of bionics is concerned with the creative
transfer of methods and structures from biology and natural systems to the engineering world. This
transfer can be described by an analogical projection of biological structures to an engineering problem
space.
On the other hand, an improper analogy can also hobble the development of an invention. Schoen (1963)
describes the case of a developer team working on synthetic fiber paintbrushes. The conceptualization of
painting as a smearing process and the comparison to natural-bristle brushes based on this idea lead to no
satisfactory results. Finally the comparison to a pumping process and the reconceptualization of the
problem in terms of this analogy paved the way to a solution. In general, the choice of an appropriate
source domain can be considered a key issue of establishing a successful analogy and so an invention
often is the end of a long series of different analogical interpretations of a situation.

Invention in Art
Initially having defined invention as the manifestation of the creative mental act, resulting amongst others
in a new artifact, concept, or the mental concretization of a conception, each artistic act is closely related
to an inventive process: The artist expresses a perception or emotion by reinventing it in terms of his
respective chosen means of artistic expression. Here, the degree to which the inventive component of the
artistic act becomes obvious can clearly vary to a high degree, featuring more prominently and accessibly
in a cubistic portrait than in a landscape painting dating back to the era of Classical Realism, and
becoming easier perceivable in a Dadaistic poem than in a naturalistic play. Nonetheless, in all these
different forms of art, a unique creative mental act finds its manifestation in an external means of
representation, giving rise to a new piece of art, and thus fulfilling the defining criteria of invention.
Invention in art does not only play a role on the microscopic scale of the individual artist, but is also
crucial on a more macroscopic level: The beginning of new styles and epochs in arts always coincides
with major inventions and inventive acts, sometimes triggered by developments outsides of arts,
occasionally arising from within arts itself. Newly available techniques and materials allow for previously
unknown means and ways of manifesting art, and new theoretical paradigms, ways of thinking and
ideologies provide the basis for new developments in arts. Examples are numerous: The invention of the
tracery marks at the start of the Gothic age, the beginning of modern painting clearly was related to the
(re)invention of the one-point perspective in the Renaissance (presumably by F. Brunelleschi), the
invention of the modern piano by B. Cristofori allowed for new ways and forms of composition heavily
influencing music as an art, the invention of motion picture by E. Muybridge sparked entirely new forms
and branches of arts, and the invention of digital media, followed by the uprising of the Internet and the
World Wide Web allowed for online art and massively distributed, decentralized art projects. Also, a
considerable number of artists were at the same time active and recognized as inventors, like R. B.
Fuller, working as an author and designer, also inventing the geodesic dome, or C. Hoberman,
architectural designer and inventor of folding toys and structures.
Concerning the role of analogy within the inventive process in arts, its impact seems more frequent on the
microscopic level of the individual artist. There, it mainly features in two ways, on a sublevel conveying
meaning and contributing to the overall impression from within the piece of art, and on a supralevel
working across several pieces of art, providing content via contextual and background effects.
 Within a piece of art, analogies may be used to provide meaning via codified metaphors and symbolic
elements grounded in the artist's and audience's environment and perception, or via free associations.
Most prominent examples for this use of analogy are probably the numerous symbolic and
iconographic elements that can be found in paintings like e.g. J. van Eyck's famous “Arnolfini
Marriage”, with cherries on a tree possibly symbolizing love, or a single lit candle alluding to the
presence of the Holy Ghost via an analogy to a sanctuary lamp.
 On the supralevel, quotations and references within a new piece of art referring to already well and
widely known earlier artworks are often to be found (providing additional information and meaning
via an analogy-based contextualization mechanism). Since the Renaissance, paintings often make
reference to scenes from Classical Antiquity, providing contextual information via the spectator's
knowledge about these classical themes and legends. Also, different compositions quote from known
and locatable sources, as e.g. Dvorak's Symphony No. 9, “From the New World”, which makes wide
use of themes from Native American music and African-American spirituals.
Nonetheless, also on the macroscopic level, mostly at turning points of the development within a certain
branch of arts, analogical mechanisms may be at work. A prominent example for such a process is the
beginning of Cubism: In its initial phase, P. Gauguin, H. Matisse and P. Picasso drew inspiration from
amongst others African, Micronesian and Native American art by transferring perceived basic principles
of these styles to a European arts context.
(III) Implications for Theory and Practice

Computer-Aided Innovation
The development of software tools for the computational support of design, e-learning, manufacturing
processes and alike is an industry in itself. A well-known example of such support systems is computer-
aided design (CAD) comprising a large number of different software systems. In comparison to such
economically important and technically already very advanced products, the research field of computer-
aided innovation (CAI) is still in a rather premature state, currently being rather an academic discipline
without reaching yet a substantial economic impact. Nevertheless, the coupling of recent advances in the
research of analogy-making with insights from cognitive science about the cognitive foundations of
innovation processes and the inspiration of classical creativity models and frameworks for innovative
problem solving is a very promising approach for the future. Such a merger of methodologies has the
potential not only to achieve theoretically new insights, but also to find practical applications that will
guide engineers and managers in developing new products and services. Due to the fact that research in
analogy-making is strongly driven by the development of computational models there is a convergence of
methodologies for CAI.

Computational Creativity
In the domain of applied Computational Creativity, analogy and analogy-related invention are relevant
and can find applications galore, supporting, enhancing, or even replacing human abilities and activities.
This can be done in two different ways: Supporting the creative activity in a purely passive tool-like way,
e.g. applying analogy-based mechanisms in search or retrieval tasks, or partly replacing human efforts in
a more active collaborative way, for instance using analogy in computing proposals for a problem
solution or interactively designing a shape.
The use of so called “references” (i.e. images of artificial or natural forms) as visual analogs and
metaphors in architectural design gives an example where a computational analogy engine can provide
support for a human user: Several databases offer sample images for such a purpose. Unfortunately, the
respective data sets are mostly indexed only with textual information describing e.g. the origin of the
image or the architect, which might not be directly relevant when searching for similar pictures to a given
image. Here, systems like Drawing Analogies by Do and Gross (1995), provide an integration of a search
mechanism for analogous shapes to an already encountered image or even to a shape provided by the
user, e.g. via a sketchpad.
The second important application area of computational analogy-making in applied Computational
Creativity are systems directly getting involved in the creative act in a computer-aided design process by
introducing new design variables. As described by Goel (1997), an example for such a system is DSSUA
(Design Supporting System Using Analogy), which addresses mechanical design problems within the
context of architectural design. DSSUA applies an analogy-based mechanism using design prototypes,
representing knowledge about familiar designs, and is even able to conduct analogical transfer for
introducing additional new variables into an initial solution to the design problem at hand.
So far of mostly academic interest is a second reading of the term Computational Creativity, in this case
denominating the attempt of modeling, simulating or replicating human creativity using a computer.
Although remarkable progress has been made, e.g. in the fields of interactive automated storytelling or
interactive drama, in automatically generated poetry or in automatically creating creative answers to
sequence completion tasks, the results still fall short when being compared to real human performance.
Closely connected to Computational Creativity are the domains of Computer-Aided and Computer-
Generated Art. Whilst Computer-Aided Art normally refers to an artistic process, where the computer is
only used as a tool, with the artwork being created by the artist, Computer-Generated Art means the
creation of artworks by using autonomous processes without direct human control. In both subfields,
analogy can play an important role: For Computer-Aided Art, the situation is similar to a Computational
Creativity support system like sketched in the Drawing Analogies example above. In the Computer-
Generated Art domain, analogy engines can find applications at the hour of creating variations of a
motive, or when computing the composition of a scenery provided an overall topic or theme is given.

Conclusions and Future Directions

As invention is the manifestation of creative mental acts, an understanding of these mental processes is
crucial to foster new inventions, be it by human inventors or by artificial systems. Analogies provide an
explanation of invention as the creative transfer of a solution to an unknown problem. A given problem
constitutes the poorly understood target domain, which can be re-conceptualized by establishing an
analogy to a better known source domain. Then, solution strategies can be developed by transferring
concepts and ideas from the source to the target. Here, in order to identify a suitable analogy, the choice
of an appropriate source domain is essential. There exists no general strategy to solve this task: Often
closely related problems are considered first, but in many cases more dissimilar domains have proven to
be the most valuable bases for analogy. However, of course, most domains do not lead to a sensible
analogy at all, and also many analogies can be misleading in the search for a solution to a given problem.
At this point, the introduction of automated support systems is promising. An analogy engine, that
operates on a knowledge base containing a large collection of domains, could propose analogies, that
exhibit a good structural compatibility to the given problem domain and that would provide a solution to
this problem. In particular, in the field of computer-aided innovation the combination of computational
techniques of analogy making and creative problem solving is a promising direction for future research.
Similarly important for future applications, although more difficult because of the lack of a controlled
underlying domain, is the usage of analogy engines in computational creativity, for example in order to
autonomously generate art by programs or to support artists by computer-aided support systems.

Cross-References
* Adaptive creativity and Innovative creativity
* Beliefs about creativity in invention
* Cognition of creativity
* Cognitive Foundations of Creativity and the Cognitive Process of Creation
* Creativity and Innovation
* Creativity between Invention and Innovation
* Invention and innovation as creative problem solving activities
* Invention vs. Discovery
* Invention, inventive learning, and innovative capacity
* Mathematical discovery
* Mental models and creative invention
* Metaphorical Reasoning and Design Creativity: consequences for practice and education
* Metaphors in Design Problem Solving: Implications for Creativity
* Models for creative invention
* Theories of creativity in invention
References
Altshuller, G. (1984), Creativity as an Exact Science. New York, NY: Gordon & Breach.

Argand, J.-R. (1813), Philosophie mathématique. essay sur une manière de représenter les quantités
imaginaires, dans les constructions géométriques, Annales de Mathématiques pures et appliquées 4, 133–
146.

Burki, L. and Cavalluci, D. (2011), Measuring the Results of Creative Acts in R & D: Literature Review
and Perspectives. In: D. Cavalluci, R. de Guio, and G. Cascini (eds.): Building Innovation Pipelines
through Computer-Aided Innovation, CAI 2011, Springer, pp.163-177.

Chomsky, N. (1957), Syntactic Structure, The Hague/Paris: Mouton.

Do, E. Y. and Gross, M. D. (1995). Drawing Analogies: finding visual references by sketching. In:
Proceedings, Association for Computer Aided Design in Architecture, 1995 National Conference, Seattle.

Goel, A. K. (1997). Design, analogy and creativity. In: IEEE Expert 12(3), May-June 1997, pp. 62-70.

Holyoak, K. J. and Thagard, P. (1995). Mental Leaps: Analogy in Creative Thought. Cambridge, Mass.:
MIT Press.

Lakoff, G. and Núñez, R. (2000). Where Mathematics Comes From: How the Embodied Mind Brings
Mathematics into Being. Basic Books.

Martinez, M, Besold, T. R., Abdel-Fattah, A., Gust, H., Schmidt, M., Krumnack, U. and Kühnberger, K.-
U. (in press). Theory Blending as a Framework for Creativity in Systems for General Intelligence. In P.
Wang and B., Goertzel (eds.): Theoretical Foundations of Artificial General Intelligence, Springer.

McGraw, G. E. (1995). Letter Spirit (part one): Emergent High-Level Perception of Gridletters Using
Fluid Concepts. PhD Thesis, Indiana University, September 1995.

Polya, G. (1954): Mathematics and Plausible Reasoning: Induction and Analogy in Mathematics,
Princeton University Press, Princeton.

Schoen, D. (1963): Displacement of Concepts, Humanities Press, New York.