Conference PaperPDF Available

Definitional Foundations for Intelligent Systems, Part II: Constructing a Definition and Examples



We posit that the lack of consensus definitions of (machine or artificial) intelligence might be affected by the lack of knowledge of conceptual analysis and other well-investigated theories. Acute contextualization of the concepts that are defined may also be an issue. Accordingly, in this two-part paper, we review some basic concepts from across research fields on how to explicate a definition. In Part I we propose 30 quality criteria for definitions that shall serve as guidelines for well-defined definitions of any concept. The quality criteria may allow for both better insights into definitions and a wider understanding of the current discourse on AI. In Part II we provide basic terminology on definitions and an iterative process to guide the construction of robust definitions by considering the quality criteria introduced in Part I. Our central goal is twofold: we want to facilitate understanding across fields and inform different stakeholders from industry, academia, legal and governments, among others, by contributing to the formal foundations on elucidating "good and robust definitions" for AI.
The 10th AC 2020 in Prague
Multidisciplinary Academic Conference
Definitional Foundations for Intelligent Systems,
Part II: Constructing a Definition and Examples
Dagmar Monett1 and Colin W. P. Lewis2
1 Computer Science Dept., Berlin School of Economics and Law, Germany, and
1, 2
We posit that the lack of consensus definitions of (machine or artificial) intelligence might be
affected by the lack of knowledge of conceptual analysis and other well-investigated theories.
Acute contextualization of the concepts that are defined may also be an issue. Accordingly, in
this two-part paper, we review some basic concepts from across research fields on how to
explicate a definition. In Part I we propose 30 quality criteria for definitions that shall serve as
guidelines for well-defined definitions of any concept. The quality criteria may allow for both
better insights into definitions and a wider understanding of the current discourse on AI. In Part
II we provide basic terminology on definitions and an iterative process to guide the construction
of robust definitions by considering the quality criteria introduced in Part I. Our central goal is
twofold: we want to facilitate understanding across fields and inform different stakeholders
from industry, academia, legal and governments, among others, by contributing to the formal
foundations on elucidating “good and robust definitions” for AI.
Keywords: Artificial Intelligence, definitions, intelligence, quality criteria
Main Conference Topic: Robotization, Engineering and Artificial Intelligence
Well-defined definitions are essential in any field. They are seen as a condition for
progress [6], help to guide research [15, 21], and show the way towards a clear target [4] or
goal of the field. However, established theories and practices that deal with the definition of
concepts are rarely considered by most practitioners. For example, a thorough conceptual
analysis of intelligence or of the concepts intelligence is based or depends on, has been
unfortunately missing in the AI community. Kelley in The Art of Reasoning [10] suggests that
a definition and the concept we define through it should provide a neutral framework for
thought and discussion “so that people on opposite sides of an issue can rely on a common
understanding of the relevant concepts in presenting their arguments and thus understand each
other.” Consequently, he states that “it is not a good idea to include controversial information
in a definition.” This is not the only desired property of a good definition, though. Several other
desirable properties or quality criteria for definitions were introduced in Part I of this paper.
There, we also introduced the problem of not having an overarching definitions of artificial
intelligence (AI) and why this has been a stumbling block in the development of the AI field.
In Part II, we introduce basic terminology on definitions as well as an iterative process for
constructing a definition. We also provide examples to clarify some of the quality criteria
introduced in Part I of the paper.
The 10th AC 2020 in Prague
Multidisciplinary Academic Conference
Basic Terminology on Definitions
In this section we first introduce the basic terminology on definitions, like the
constituents of a definition, its types and purpose, among other related subjects. The section’s
aim is to raise awareness of the importance of these subjects, long overlooked by most
stakeholders although essential to the definitional foundations of any field.
What is a Definition?
A definition “states a thing’s essence,” it “expresses the nature of a thing and its
substance.” These were Aristotle’s words in his Posterior Analytics [1] written in 350 B.C.E.
Since then, a myriad of definitions of what a definition is has been proposed, but almost all of
them converge to the same idea: the purpose of a definition is to state, express, explain, or
specify what a thing is. According to the Cambridge Dictionary, a definition is not only “a
statement that explains the meaning of a word or phrase,” i.e. its essence, its substance, but
also “a description of the features and limits of something,” i.e. its characteristics and scope.
Definitions can be defined by expressing the meaning of the thing to be defined in very different
ways: there are definitions that do it by using other constructs where the meaning is known;
these are the lexical definitions when concepts from a lexicon or dictionary are used. Ostensive
definitions exemplify things to which the expression applies because examples are pointed out.
And stipulative definitions express the meaning by proposing how the language should be used.
These three broad categories, lexical, ostensive, and stipulative definitions, have received
different designations with time and have been subdivided into further subcategories.
How Should a Definition be Defined?
Definitions enable people to understand a concept or theory; thus, providing conceptual
clarity on how definitions should be defined is, although mostly overlooked, an essential
activity prior to the use of the concept or theory. According to Cassidy [3] after his analysis of
several Aristotelian works on definitions, a concept may have many definitions and how it is
defined “is a function of the language in which it is stated,” i.e. “the way we talk about things
is a function of the things we talk about.” Furthermore, a definition is a “verbal formulae of
selectively grouped data taken from the knower’s experience” [3]. Hence, there is no unique
way of defining a concept, definitions may be inherently contextual, and they would depend
on both the people that define them and their area of knowledge.
A definition should have the form of “an equivalence” [20], of “a two-part equivalence”
[18], which on the one hand contains the thing to be defined, the definiendum, and on the other
hand the expression that defines it, the definiens. Ideally, the definiens follows the definiendum
in order of appearance and both are constructed in such a way that certain Rules of Definitions
[10, 20] are fulfilled. The definiendum, for example, should be a short, grammatically simple
expression that specifies the concept that is being defined. Similarly, the definiens should fulfill
certain rules that guarantee, for example, that the concepts used in its expression have a
meaning that is previously known or that has already been defined: “[i]t may be a sentential
function of an arbitrary structure, but containing only those constants whose meaning is
understood; that is, the meaning of each of those constants either is immediately obvious or
has already been explained” [20]. Furthermore, it should not be too broad or too narrow,
circular, nor use negative terms unnecessarily, ambiguity or vagueness, nor obscure or
metaphorical language [10].
In his empirical work on Logic, Mercier [14] states that to define “is to say what a thing
is.” One way of stating the meaning of the definiendum is by referring to its genus and
The 10th AC 2020 in Prague
Multidisciplinary Academic Conference
differentia as part of the definiens, like in Aristotelian definitions. The genus or kind is the
broader category the definiendum belongs to. For example, in Gottfredson’s definition
“Intelligence is a very general mental capability that [...] involves the ability to reason, plan,
solve problems, think abstractly, comprehend complex ideas, learn quickly and learn from
experience” [7], mental capability is the genus of intelligence, i.e. the broader concept. The
differentia, however, distinguishes the concept from other definienda, in that it specifies the
properties or characteristics that differentiate the definiendum from other concepts from the
same genus, as it is expressed in the rest of Gottfredson’s example.
A definition could have different functions and be “relative to various purposes” [22].
It can clarify the boundaries of a concept, the relationships among concepts, and it can also
provide a summary statement about the referents of a concept, i.e. it can highlight its essence
[1, 10]. According to Kelley [10], “[a] good definition condenses the knowledge we have about
the referents of a concept, giving us just the highlights, the key points, the essence.”
Conceptual Analysis
One of the main activities philosophers deal with is conceptual analysis. As pointed out
by Aaron Sloman [19], conceptual analysis is “required for improving our understanding not
of the physical world itself, but of how people of various ages and cultures think about the
world.” On the other hand, our understanding of the world builds on certain assumptions
without it would not be possible to construct or define most common concepts we know [5].
When we define a concept, we fix it, we isolate it from the rest [22], thus how we define it
matters. According to [13], conceptual analysis “specifies a set of conditions that are
individually necessary and jointly sufficient for the application of [a] concept.” As such,
conceptual analysis encompasses a series of strategies and methods that seek to analyze and
describe concepts, phenomena, and theories. Some of these strategies are extensively detailed
in [19] and summarized in what follows. They may be a necessary step that must be considered
before constructing a definition of a concept, in parallel to the process of defining it, or even
after doing that, for instance when a definition must be refined or improved to cope with
advances in a particular field of study. Some strategies presented in [19] for analyzing a given
concept are:1
Collecting descriptions of both varied instances and non-instances of the concept, and
comparing them. This would help to clarify the specificity with which the concept is or
should be defined.
Criticizing and extending already existent definitions, e.g. definitions given in
dictionaries or by other practitioners in the field. A thorough revision of the literature
should avoid unnecessary re-definitions of an already defined terminology.
Collecting examples of the concept and related words that may illustrate one concept but
not the others; also collecting examples of their use and how to teach the concept to other
people. This might help to understand the concept and its definition by others, as it will
be discussed later in this paper.
Asking what the role of the concept is in a given culture or context, which should its
purpose be, whether it would guide research in a given field and how. This could
determine the type of the definition and the way it is constructed.
1 We refer the reader to the cited work for a detailed description of these and other strategies.
The 10th AC 2020 in Prague
Multidisciplinary Academic Conference
Listing ways in which the truth or falsity of statements involving the concept in question
can be tested. This may allow for the conceptual clarity to which a good definition ought
to contribute.
How to Construct a Good Definition?
The Rules of Definitions [10, 20] introduced above consider a set of properties that a
definition should meet in order to be considered a good definition. Kelley [10] argues, for
example, that if a definition includes a genus and a differentia that are both not too broad and
not too narrow, and that state the essential attributes of the concept referents, then it is quite
probable that it will also satisfy other rules from the set he proposes which is comprised by six
rules altogether. One way of searching for such a combination of genus and differentia, which
is the classic way to define a concept, is by following Plato’s method of definition known as
Diairesis or method of division, developed in [17] and later refined by Aristotle in his Posterior
Analytics [1]. The idea is that a definition should include all essential elements that make it
both exclusive and comprehensive (see quality criteria 14 and 15 from the list presented in Part
I of this paper) by successively considering the similarities and distinctive properties of these
elements with regard to other instances of the same category, thereby avoiding both the
omission of important characteristics and the inclusion of others that are not essential to the
definition. This way of constructing a definition is consistent with the first three rules of
definitions presented in [10] and commented at the beginning of this section. However, they
are not the only criteria that are desirable for having a good definition, as we have already
discussed in Part I of this paper. In addition to this, there is no unique way of defining a concept
(see a compendium in [5]). We suggest to construct a definition as part of an iterative process.
The one presented in what follows extends the process presented in [10], in that many more
quality criteria are included and a few of them are strengthened after considering the results
discussed here and in [16]. Our iterative process is depicted in Figure 1. Such an iterative
process could be preceded by a systematic conceptual analysis as suggested in section about
conceptual analysis. It consists of the following steps:2
(i) Use a two-part equivalence to set the form of the definition: define the definiendum as
short and simple as possible (for example, start the definition with “Intelligence is ...”).
Define the definiens with a genus-differentia expression that states the “what,” the “how,”
and the “why” of the concept that is being defined (see quality criteria 1 to 4 from Part I
of this paper) as short and simple as possible. In particular, once the genus or broader
category is found, “choose a differentia that distinguishes [the concept] from other
concepts in the same genus [i.e. which is the genus of transparency and which other
concepts belong to the same category?]. If there is more than one distinguishing
attribute, choose the most essential one,” as suggested in [10], because it would be easier
to explain what makes that attribute special, distinguishable from the rest. A previous
conceptual analysis may shed light in this endeavor.
(ii) State the differentia in an ostensive way (see the section on basic terminology, as well as
quality criterion number 5 from Part I of this paper). For example, when defining the
concept of human agency, include capabilities or functions that indicate agency (i.e.
exemplify the “what” in order to understand it better). Do not make the definition too
ostensive nor too narrow, however.
2 The Defintly App, available at and with provisional password agisiorg, contains most of
the content presented in this two-part paper and assists users when defining concepts in any field.
The 10th AC 2020 in Prague
Multidisciplinary Academic Conference
(iii) Go through the list of quality criteria for definitions (see Part I of this paper) and evaluate
whether each of them fits the functions of the definition (see section on basic terminology
above) and in which context and extent.
(iv) Add (resp. remove) content to (resp. from) the definiens to meet the quality criteria. Seek
for simplicity but sought for as many fulfilled quality criteria as possible.
(v) Adapt the wording and remove unnecessary expressions.
(vi) Iterate the steps i) to v) as long as it is needed. In particular, take a look at the language
and use it elegantly, thereby ensuring that the already-met quality criteria are still
fulfilled, i.e. that they are not violated by occasionally changing the language.
(vii) Stop the process when no more improvements are necessary.
Figure 1: Iterative process to construct a definition
Constructing a definition that satisfies all quality criteria might not be possible. Even
keeping all of them in mind when following the iterative process might not be straightforward
at the beginning. With time and practice, however, it would be more intuitive and simpler to
define concepts that follow these principles. The advantages in doing so would be considerably
positive. Each and every person involved would benefit from a well-defined definition: not
only the author(s) of the definition but also the reader(s) and people that use it. Hence, both the
community and the scientific discourse would profit likewise and advance further, because it
would be easier to discard less expressive definitions and, thus, easier to reach consensus.
The only purpose of this section is to exemplify which quality criteria from the list
presented in Part I of this paper are met or not by a definition. It is not in the scope of this paper
to construct a new definition of intelligence, since it would require at least an added, extensive
analysis of the several already existent definitions of intelligence from the literature (refer to
[11] for a compendium) as already discussed in the section on conceptual analysis. Instead, a
selection of examples of new, suggested definitions from the survey presented in [16] are
presented below. Their most relevant characteristics are analyzed and discussed.
“Intelligence is the ability to do the right thing at the right time given a dynamic
environment (that is, a shifting landscape of `right times’ requiring more `right things’).
The 10th AC 2020 in Prague
Multidisciplinary Academic Conference
[Artificial Intelligence] is intelligence constructed deliberately as an artefact of a
culture.” Joanna Bryson, University of Bath and University of Princeton
Bryson’s definition defines the “what” of the definition and it makes an explicit
distinction between human and artificial intelligence. It is not ostensive (i.e. it doesn’t include
examples), but it could be considered operational since it exemplifies “how” to reach
intelligence, i.e. by “[doing] the right thing.” This definition is short, affirmative, and general.
If separated from the whole, the second part of the definition would contain a vicious circle.
“To my mind, intelligence (whether we’re speaking of a human or a machine) is the
ability to put one’s finger on the essence of situations that one faces, and to do so
reasonably rapidly. Or a bit more verbosely, … ‘intelligence is the art of rapid and
reliable gist-finding, crux-spotting, bull’s-eye hitting, nub-striking, essence-pinpointing.
It is the art of, when one is facing a new situation, swiftly and surely homing in on an
insightful precedent (or family of precedents) stored in the recesses of one’s memory.
That, no more and no less, is what it means to isolate the crux of a new situation. And
this is nothing but the ability to find close analogues, which is to say, the ability to come
up with strong and useful analogies.’” Douglas Hofstadter, Indiana University, also
citing [9]
Hofstadter’s definition defines the “what” and is ostensive: it exemplifies cognitive
abilities that indicate intelligence. It also defines the “how” and includes examples of
behavioral functions to reach intelligence, thus it is operational. It is not a short definition, but
it is affirmative, comprehensive, and useful. Furthermore, it is an intellectually elegant
definition. However, it might not be a clear definition for some audiences. Finally, let’s analyze
some definitions from the literature:
“Intelligence is a very general mental capability that, among other things, involves the
ability to reason, plan, solve problems, think abstractly, comprehend complex ideas,
learn quickly and learn from experience. It is not merely book learning, a narrow
academic skill, or test-taking smarts. Rather it reflects a broader and deeper capability
for comprehending our surroundings -- `catching on,’ `making sense’ of things, or
`figuring out’ what to do.” [7]
Gottfredson’s definition, which was already used as an example in the section on basic
terminology above, was the most agreed upon definition of intelligence in [16]. Of the 567
people that participated in the survey,3 a total of 356 respondents agreed (38.1%) or strongly
agreed (24.7%) with this definition, representing about two-thirds of positive agreement. It is
also a widely accepted definition among experts in intelligence and allied fields [8], such as
neuroscientists and psychologists. Linda Gottfredson is a psychologist. It should be mentioned
that more than 70% of all participants in the survey are computer scientists, engineers,
mathematicians, and physicists. Is Gottfredson’s definition a good definition of intelligence?
One of its most distinguishing characteristics is that it exemplifies several cognitive abilities
when defining the “what,” i.e. when defining intelligence. In fact, it was the most ostensive
definition from the survey, which presented a total of 18 definitions of intelligence from the
literature to agree upon. It can be verified that it fulfills almost all 30 but three quality criteria
(no. 17, 18, and 30). This is very interesting since other definitions explicitly avoid the
3 The AGISI survey was closed in July 2019. The results included here are considered to be final. Partial results
are presented in [16].
The 10th AC 2020 in Prague
Multidisciplinary Academic Conference
specification or exemplification of capabilities and are therefore considered more abstract and
general by their authors, thus better for defining intelligence. That is the case of Legg and
Hutter’s definition of machine intelligence [12]: “Intelligence measures an agent’s ability to
achieve goals in a wide range of environments.” The authors analyze several examples of
definitions in detail and claim that intelligence is the effect of capacities and not the result of
having a set of them. Legg and Hutter mostly refer to semantic aspects, i.e. to which extent and
how the analyzed definitions define intelligence or not, thereby justifying why their own
proposed definition of intelligence might be more appropriate. Furthermore, they argue that it
is very difficult to specify the capacities that might be needed and that they would depend on
the concrete context at hand. It seems contextual definitions are too specific for some authors.
However, ostensive definitions are more preferred by researchers and practitioners: Legg and
Hutter’s definition had only a 48.7% of total positive agreement from the respondents to the
survey (276 respondents agreed (37.4%) or strongly agreed (11.3%) with that definition) and
it was ranked fourth place overall. Other definitions, Gottfredson’s included, were much more
favorably ranked by respondents. A commonality is that they are ostensive.
If we look back at the examples presented in this section, not all of them follow the
Aristotelian construct. Moreover, not only one genus but different genera were used by the
authors to define the broader category intelligence may belong to. Which of them is the most
appropriate level of abstraction might be in itself a very controversial topic. Ideally, a good
definition satisfies many of the 30 quality criteria for definitions presented in Part I of this
Part II of this paper introduced definitional foundations for definitions. These included a
process for constructing good definitions of concepts, from scratch, in particular, the concept
of intelligence. Examples were also provided to illustrate some of the quality criteria that were
introduced in Part I. We do not recommend a technical perspective for the analysis of these
topics; our wish would be the AI community being more knowledgeable about what other non-
technical disciplines have considered so far. A similar analysis as the one presented here for
the concept of intelligence could be applied to the definition of those concepts that are having
a crucial role in establishing the guidelines and foundations for AI. We hope our work can
contribute to informing different stakeholders from industry, academia, legal and governments,
among others, on how to do such an analysis.
We would like to thank Aaron Sloman for his enlightening suggestions about conceptual
analysis and its relation to definitions.
The 10th AC 2020 in Prague
Multidisciplinary Academic Conference
[1] Aristotle (1901). Aristotle’s Posterior Analytics. Oxford: Blackwell.
[2] Beckett, C. (2019). New powers, new responsibilities: A global survey of journalism and
artificial intelligence. Department of Media and Communications, The London School
of Economics and Political Science.
[3] Cassidy, J. R. (1967) Aristotle on definitions. Southern Journal of Philosophy, 5, 110
[4] Chollet, F. (2019). The measure of intelligence. arXiv:1911.01547 [cs.AI].
[5] Cregan, A. M. (2005). Towards a science of definition. In T. Meyerand & M. Orgun
(eds.), Proceedings of the Australasian Ontology Workshop, AOW 2005, 58, 2532,
Australian Computer Society, Inc.
[6] De Boeck, P. (2013). Intelligence, where to look, where to go? Journal of Intelligence,
1:524, 2013.
[7] Gottfredson, L. S. (1997). Mainstream science on intelligence: An editorial with 52
signatories, history, and bibliography. Intelligence, 24, pp. 1323.
[8] Haier, R. J. (2017). The Neuroscience of Intelligence. New York, NY: Cambridge
University Press.
[9] Hofstadter, D., & Sander, E. (2013). Surfaces and Essences: Analogy as the Fuel and
Fire of Thinking. New York, NY: Basic Books.
[10] Kelley, D. (2014). The Art of Reasoning: An Introduction to Logic and Critical Thinking.
Fourth edition, New York, NY: W.W. Norton & Company.
[11] Legg, S., & Hutter, M. (2007). A collection of definitions of intelligence. In B. Goertzel
and P. Wang (eds.), Advances in Artificial General Intelligence: Concepts, Architectures
and Algorithms, 157, 1724, UK: IOS Press.
[12] Legg, S., & Hutter, M. (2007). Universal Intelligence: A Definition of Machine
Intelligence. Minds and Machines, 17, 391444.
[13] Margolis, E., & Laurence, S. (2019). Concepts. The Stanford Encyclopedia of
[14] Mercier, C. (1912). Elements of Logic. Third edition, New York, NY: The Manhattanville
[15] Mikolov, T. (2020). Why is defining artificial intelligence important? Journal of
Artificial General Intelligence, 11(2), 5051.
[16] Monett, D., & Lewis, C. W. P. (2018). Getting clarity by defining artificial intelligence
A Survey. In V. C. Müller (Ed.), Philosophy and Theory of Artificial Intelligence 2017,
SAPERE 44, 212214, Berlin: Springer.
[17] Plato (1892). The Dialogues of Plato. Third edition, vol. 4, Oxford University Press.
[18] Random, R. A. (1958). Intelligence as a science. Studies in Intelligence, 2, 76.
[19] Sloman, A. (2019). The Computer Revolution In Philosophy: Philosophy, science and
models of mind. Revised, online edition, Sussex: Harvester Press.
[20] Tarski, A. (1994). Introduction to Logic and to the Methodology of Deductive Sciences.
Fourth edition, New York: Oxford University Press.
[21] Wang, P. (2019). On defining artificial intelligence. Journal of Artificial General
Intelligence, 10, 37.
[22] Watts, A. W. (1951). The Wisdom of Insecurity: A message for an age of anxiety. Second
edition, New York: Vintage Books.
The 10th AC 2020 in Prague
Multidisciplinary Academic Conference
Brief biographies of the authors
Dagmar Monett
Prof. Dr. Computer Science (Artificial Intelligence, Software Engineering) with 30+
years of research and teaching experience. Co-founder of the AGI Sentinel Initiative,, dedicated to understanding intelligence in order to build beneficial AI and to
educate society’s stakeholders across disciplines. AI expert at Ms.AI, “Artificial Intelligence
for and with Women.”
Colin W. P. Lewis
Adjunct Professor and international business and policy advisor, specialized in the
social and economic impact of technological change. With specific knowledge of the evolving
conditions of technology for productivity growth, development, labor, (in)equality, and
competitiveness. Part of his work focuses on understanding human intelligence and creating
solutions for Artificial Intelligence.
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
This article systematically analyzes the problem of defining “artificial intelligence.” It starts by pointing out that a definition influences the path of the research, then establishes four criteria of a good working definition of a notion: being similar to its common usage, drawing a sharp boundary, leading to fruitful research, and as simple as possible. According to these criteria, the representative definitions in the field are analyzed. A new definition is proposed, according to it intelligence means “adaptation with insufficient knowledge and resources.” The implications of this definition are discussed, and it is compared with the other definitions. It is claimed that this definition sheds light on the solution of many existing problems and sets a sound foundation for the field.
Full-text available
The first issue of the Journal of Intelligence is devoted to a discussion based on the following two questions: 1. What are the most important scientific issues in the domain of human intelligence? 2. What are the most promising new ideas and approaches in the study of human intelligence? [...]
Full-text available
Analogy is the core of all thinking. This is the simple but unorthodox premise that Pulitzer Prize–winning author Douglas Hofstadter and French psychologist Emmanuel Sander defend in their new work. Hofstadter has been grappling with the mysteries of human thought for over thirty years. Now, with his trademark wit and special talent for making complex ideas vivid, he has partnered with Sander to put forth a highly novel perspective on cognition. We are constantly faced with a swirling and intermingling multitude of ill-defined situations. Our brain’s job is to try to make sense of this unpredictable, swarming chaos of stimuli. How does it do so? The ceaseless hail of input triggers analogies galore, helping us to pinpoint the essence of what is going on. Often this means the spontaneous evocation of words, sometimes idioms, sometimes the triggering of nameless, long-buried memories. Why did two-year-old Camille proudly exclaim, “I undressed the banana!”? Why do people who hear a story often blurt out, “Exactly the same thing happened to me!” when it was a completely different event? How do we recognize an aggressive driver from a split-second glance in our rearview mirror? What in a friend’s remark triggers the offhand reply, “That’s just sour grapes”? What did Albert Einstein see that made him suspect that light consists of particles when a century of research had driven the final nail in the coffin of that long-dead idea? The answer to all these questions, of course, is analogy-making—the meat and potatoes, the heart and soul, the fuel and fire, the gist and the crux, the lifeblood and the wellsprings of thought. Analogy-making, far from happening at rare intervals, occurs at all moments, defining thinking from top to toe, from the tiniest and most fleeting thoughts to the most creative scientific insights. Like Gödel, Escher, Bach before it, Surfaces and Essences will profoundly enrich our understanding of our own minds. By plunging the reader into an extraordinary variety of colorful situations involving language, thought, and memory, by revealing bit by bit the constantly churning cognitive mechanisms normally completely hidden from view, and by discovering in them one central, invariant core—the incessant, unconscious quest for strong analogical links to past experiences—this book puts forth a radical and deeply surprising new vision of the act of thinking.
Full-text available
A fundamental problem in artificial intelligence is that nobody really knows what intelligence is. The problem is especially acute when we need to consider artificial systems which are significantly different to humans. In this paper we approach this problem in the following way: We take a number of well known informal definitions of human intelligence that have been given by experts, and extract their essential features. These are then mathematically formalised to produce a general measure of intelligence for arbitrary machines. We believe that this equation formally captures the concept of machine intelligence in the broadest reasonable sense. We then show how this formal definition is related to the theory of universal optimal learning agents. Finally, we survey the many other tests and definitions of intelligence that have been proposed for machines.
Full-text available
This paper is a survey of a large number of informal definitions of ``intelligence'' that the authors have collected over the years. Naturally, compiling a complete list would be impossible as many definitions of intelligence are buried deep inside articles and books. Nevertheless, the 70-odd definitions presented here are, to the authors' knowledge, the largest and most well referenced collection there is.
Traducción del griego Reimpresión en 1987 Incluye bibliografía Euthyphro -- Apology -- Crito -- Meno -- Phaedo