And yet...Lakoff uses the method of Ayn Rand to build his theory

Condensing Experience:

Hierarchy, Metaphor, and Randian Concept Formation in Lakoff and Johnson

A Comparative Essay

September 2026

I. The Question

In Introduction to Objectivist Epistemology, Ayn Rand treats language as an epistemological instrument. A concept condenses the essential features of a class of existents; a word converts that condensation into a single mental concrete; a definition abbreviates the essentials still further; and isolation pulls the relevant units out of the perceptual stream so that the rest of the stream may pass into forgetfulness.[1] The governing problem is cognitive economy. A consciousness of limited range cannot hold the perceptual flood as a flood. It must reduce that flood to units it can keep, compare, and use.

George Lakoff and Mark Johnson, in Metaphors We Live By and in Lakoff’s Women, Fire, and Dangerous Things, begin from a recognizably similar pressure. Experience does not arrive already carved into bounded objects with shared essential properties. A finite, embodied mind must pick out parts of that experience, treat them as discrete entities, and impose structure sufficient for reference, grouping, quantification, and action.[2] Two instruments do most of the work in their account: a hierarchy of categories, from basic-level cuts up through superordinates and out along radial extensions, and conceptual metaphor, which builds and extends that hierarchy by projecting the structure of one domain onto another.

An earlier comparison could stop at a qualified verdict. Lakoff and Johnson share Rand’s claim that concepts make untamed experience manageable. They do not share her claim that the mechanism is grouping by essence and the omission of measurements. The present essay takes a further step. It asks whether the methods Lakoff and Johnson themselves use—their characterization of linguistic hierarchy, and their use of metaphor to construct and extend that hierarchy—have essential characteristics, and whether those characteristics are an instance of the very operation Rand describes: the isolation of a commensurate distinguishing feature and the omission of its particular measurements. The answer defended here is that they do, and that they are, with a restriction that matters. At the level of theoretical concept-formation, Lakoff and Johnson practice a Randian method. At the level of some of the folk categories they describe, especially those structured by chaining and family resemblance, the fit is only partial. The interesting result is not that the two theories collapse into one another. It is that Lakoff and Johnson’s anti-essentialist object-theory is framed in an essentialist theoretical vocabulary, and that the vocabulary is forced on them by the same cognitive requirement Rand names unit-economy.

II. Rand: Quantitative Similarity and Unit-Economy

Rand’s solution to the problem of universals is a quasi-mathematical analysis of similarity. Similarity is “the relationship between two or more existents which possess the same characteristic(s), but in different measure or degree.”[3] The grasp of similarity requires a contrast object that differs along the same scale. That shared scale is the conceptual common denominator.[4] A concept is “a mental integration of two or more units possessing the same distinguishing characteristic(s), with their particular measurements omitted.”[5] The measurements are not treated as nonexistent. They must exist in some quantity and may exist in any quantity. What is omitted is their particularity, so that the concept may stand for an open-ended class.

Four operations complete the process. Isolation selects related items from the perceptual field. Integration blends them into a new mental entity used as one unit of thought. A word gives that entity “the form of a single, specific, perceptual concrete.”[6] A definition states the essential attributes of the concept, where essence is epistemological: the characteristic that, in the present context of knowledge, explains the most and distinguishes the units from everything else so far identified.[7]

These operations are required by unit-economy. Consciousness cannot perceive everything at once, nor hold an unlimited number of units in focal awareness. “The essence, therefore, of man’s incomparable cognitive power is the ability to reduce a vast amount of information to a minimal number of units.”[8] Concepts “represent condensations of knowledge.”[9] Their primary purpose, and the primary purpose of language, is not communication but cognition: “to keep order in man’s mind and enable him to think.”[10]

Two features of this account will carry the later argument. First, “quantitative” does not mean that every concept is a number. It means that similarity is sameness of characteristic across difference of degree. Length, color, speed, intelligence, justice, and causal power can all serve as common denominators if they can be held as the same attribute in varying measure. Second, the method is not confined to first-level perceptual concepts. Abstraction from abstractions repeats the same operation on earlier concepts. A scientific or philosophical concept—metaphor, hierarchy, prototype—is, if valid, formed in the same way: isolate the units, retain the distinguishing characteristic, omit the particular measurements.

III. Lakoff and Johnson: Hierarchy and Metaphor as Methods

Lakoff and Johnson open Metaphors We Live By with the claim that “our ordinary conceptual system, in terms of which we both think and act, is fundamentally metaphorical in nature.” The essence of metaphor, they write, is “understanding and experiencing one kind of thing in terms of another.”[11] Experience that is not already a bounded object is unitized by ontological metaphor: “Understanding our experiences in terms of objects and substances allows us to pick out parts of our experience and treat them as discrete entities or substances of a uniform kind.”[12] Categorization then highlights some properties, downplays others, and hides still others.[13]

The hierarchy in their theory has several layers, and it is easy to conflate them. There is the taxonomic hierarchy of basic-level, superordinate, and subordinate categories, taken from Eleanor Rosch: chair rather than furniture, dog rather than mammal. Basic-level categories are learned earlier, recognized faster, and tied to a characteristic shape and a characteristic motor program.[14] There is the radial hierarchy of a category organized around a central case, with extensions licensed by metaphor, metonymy, and cultural association. Mother in this sense is not a flat set. It is a network with a birth-and-nurture center and a periphery of adoptive, step, surrogate, and genetic cases.[15] There is the image-schematic structuring of categories themselves as containers, part-whole assemblies, and center-periphery organizations. And there is the metaphorical scaffolding that lets an abstract or poorly bounded domain borrow the hierarchy of a concrete one: theories have foundations and collapse; arguments have positions and lines of attack; purposes have destinations.

Metaphor is not, in this program, an ornament applied after the hierarchy is in place. It is one of the operations by which the hierarchy is built. A source domain that is already structured—war, journeys, buildings, verticality, containment—supplies relations of inclusion, degree, opposition, and goal that the target does not present as ready-made. The mapping is partial. Some structure is carried; the rest is left behind. The result is a usable hierarchy in a domain that would otherwise remain a flood.

Lakoff’s emblem of a non-classical category remains the Dyirbal classifier balan, which groups women, fire, water, and dangerous things by motivated chaining rather than by a single defining attribute.[16] The polemical frame of both books is a critique of “objectivism”: the world pre-sorted into classical kinds, meaning as a correspondence between disembodied symbols and that sorting, metaphor as cognitively idle.[17] That target is not identical with Rand’s Objectivism. Rand denies that knowledge is a passive copy; she treats essence as epistemological; she treats consciousness as an active process.[18] The verbal collision is therefore partly accidental. The question that remains is whether their own method of isolating “basic level,” “radial category,” and “conceptual metaphor” can do without the operation they officially reject.

IV. The Method Has Essential Characteristics

A theory that denies essential characteristics to ordinary categories may still form its own theoretical concepts by isolating essential characteristics. That is not a debating trick. It is what any systematic description must do if it is to be a description of something rather than a list of episodes. Lakoff and Johnson do not report an unstructured heap of linguistic facts. They isolate kinds. Those kinds have definitions, even when the definitions are given as characterizations rather than as genus-and-differentia formulas. The question is whether the isolation is an instance of Rand’s method.[19]

Forming the concept metaphor

To write a book about conceptual metaphor is already to have formed the concept conceptual metaphor. The units are particular mappings: ARGUMENT IS WAR, TIME IS MONEY, THEORIES ARE BUILDINGS, MORE IS UP, PURPOSES ARE DESTINATIONS. What is the same across those units, in different measure and in different domains, is a cross-domain projection of inferential structure: one kind of thing is understood in terms of another, systematically, not as a one-off resemblance. What is omitted are the particular source domains, the particular targets, the particular lexical realizations, and the particular cultural settings. That is measurement-omission. The word metaphor is the single mental concrete that holds the open class. The statement that metaphor is understanding one kind of thing in terms of another is a definition by essentials.[20]

Lakoff and Johnson themselves insist that the mapping is partial.[21] Partiality is the counterpart of omission. Not every feature of war is a feature of argument; not every feature of a journey is a feature of a life. The dimensions that transfer—opposition, goal, progress, obstruction—are the conceptual common denominator of the metaphor. The dimensions that do not transfer are the measurements left behind. Rand’s language was built for attributes of entities, but the operation is the same when the “characteristic” is a relational structure rather than a magnitude.[22]

Forming the concepts of hierarchy

The same pattern holds for the hierarchical apparatus. Basic-level category isolates a cut in a taxonomy distinguished by gestalt perception, characteristic motor interaction, and a single mental image. The particular contents—chair, dog, tree, cup—are measurements omitted. Superordinate and subordinate name positions on a scale of inclusiveness and of imagistic specificity. Prototype isolates a central case relative to which other members stand at some degree of distance. The typicality ratings that support the theory are already quantitative: goodness-of-example is the same characteristic in different measure.[23] Radial category isolates a network with a center and licensed extensions. Idealized cognitive model isolates a structured cluster of defaults against which instances can be compared. In each case a distinguishing structure is retained and the particular lexical filling is dropped. Without that condensation, Women, Fire, and Dangerous Things would be an album of examples rather than a theory of categorization.

Hierarchy, moreover, is itself understood by Lakoff in image-schematic terms: categories as containers, taxonomies as part-whole and up-down structures, radial categories as center and periphery.[24] Those schemas are scales. A member is more or less central. A category is more or less inclusive. A concept is more or less schematically specified. Rand’s formula—same characteristic, different measure or degree—describes exactly that kind of order. The verticality of a taxonomy and the centrality of a prototype are not metaphors added after the fact to a non-quantitative grouping. They are the grouping’s quantitative form.

Metaphor as a builder of hierarchy

The further claim is that metaphor is not only itself a Randian concept, but an operation that constructs hierarchy by the same means. A source domain that is already articulated into parts, degrees, and inclusions supplies a common denominator that the target lacks, or possesses only implicitly. MORE IS UP does not invent quantity. It makes quantity holdable by aligning it with a bodily scale whose degrees are already perceptually available.[25] ARGUMENT IS WAR does not invent opposition. It makes the structure of an exchange holdable by aligning it with a domain in which opposition, goal, and relative position already come in degrees. The hierarchy in the target—central moves and peripheral ones, stronger and weaker positions, progress and setback—is the source hierarchy with particular martial measurements omitted.

Rand describes higher-order concepts as abstractions from abstractions, still answerable to a chain that terminates in percepts.[26] Metaphor, read this way, is one route such abstraction can take. The source domain is closer to the sensorimotor ground; the target is more abstract; the mapping retains structure and omits measure. Nothing in that description requires that the target already display a classical essence of its own. It requires only that source and target be commensurable along the dimensions the mapping keeps. Commensurability is Rand’s condition for similarity. Metaphor satisfies it by construction: a mapping that found no common denominator would not be a mapping. It would be a pun.

V. How Far the Identification Goes

If the preceding is right, Lakoff and Johnson’s method is an example of Randian concept-formation in two senses. First, their theoretical kinds are formed by isolating distinguishing structures and omitting particular fillings. Second, the object-level operations they attribute to speakers—cutting a basic level, centering a prototype, projecting a source hierarchy onto a target—are themselves descriptions of quantitative similarity, once “quantity” is allowed to include degree of typicality, degree of schematicity, and degree of structural correspondence.

The identification is not total. Three limits keep the theories from simply coinciding.

Chaining without a single denominator

Family resemblance and motivated chaining are designed to link A to B and B to C where A and C share no one characteristic that can serve as a common denominator for the whole class. Wittgenstein’s games and Dixon’s balan are the standard cases.[27] A Randian reconstruction must say one of three things: that the classifier is not a concept in the strict sense, but a conventional grouping; that some overlooked commensurate attribute has been missed; or that the chain is a series of local concepts rather than one concept. Lakoff takes the data to show something else: that a public, projectable category need not be the condensation of one essential feature. On that point the object-level theories still diverge. What does not diverge is the theorist’s own concept radial category or motivated chain. Those are well-formed. They have a distinguishing structure—center plus licensed links—and they omit the particular links.

Multiple centers and revised definitions

Radial models may have more than one center, and they may be extended without a commitment that every member shares the attributes of the original prototype.[28] Rand can revise a definition when the context of knowledge requires it, and she can treat borderline cases as problems of application.[29] She cannot, without changing theories, treat a concept as a network with several centers and no single defining measurement. The strongest friendly reading is that each center is itself a first-level condensation, and that the word covering the network is a higher-order grouping whose common denominator is a looser relational structure—kinship-and-care, in the case of mother—held across different measures of biological, legal, and social involvement. That reading is available. It is not the reading Lakoff officially wants, because it restores an essence at the next level up.

What they officially say about definition

Lakoff and Johnson reject the classical theory of definition as a list of necessary and sufficient conditions and treat understanding as organized by prototypes, metaphors, and experiential gestalts.[30] Rand’s definitions are not that classical theory. Essence for her is contextual and epistemological, and a concept includes yet-to-be-discovered characteristics of its referents.[31] Even so, she holds that a unit is in a class or it is not once the definition is fixed. The present argument does not require that Lakoff and Johnson accept that bivalence for every folk category. It requires only that their own theoretical categories behave that way. A given mapping either is or is not an instance of conceptual metaphor, in the sense they have isolated. A given lexical item either is or is not being used as a basic-level term in a given contrast set. If those determinations were indefinitely graded all the way up, the books could not state theses.

VI. Worked Examples

Table

For Rand, perceptually given tables are isolated by contrast, integrated by omitted measurements of size and style, and held by the word table. For Lakoff, table is a basic-level category: characteristic shape, characteristic interaction, a single mental image. The two descriptions now look like one operation described at two grains. The basic-level privileges a cut at which measurement-omission is easiest, because the distinguishing structure—support at a usable height for human action—is perceptually and motorically salient. Superordinate furniture repeats the operation at a higher level of schematicity. The hierarchy is not a rival to quantitative similarity. It is similarity stratified by degree of omitted particularity.

Inflation

Rising prices across many goods are not a single perceptible object. Rand forms the concept by integrating instances of a rising general price level and omitting particular rates. Lakoff and Johnson treat ordinary talk of inflation as an ontological metaphor: a process recast as an entity.[32] The metaphor is the condensation. What is retained is boundedness and causal efficacy, the structure of an object. What is omitted is the particular temporal unfolding, the particular basket of goods, the particular institutional setting. INFLATION IS AN ENTITY is measurement-omission applied to a process that would otherwise remain too fluid to hold as one unit.

Argument

ARGUMENT IS WAR highlights attack, defense, and strategy and hides the cooperative search for truth.[33] Read as a rival to Rand, the metaphor looks like a distortion of what arguments essentially are. Read as an instance of her method, the metaphor isolates a commensurate structure—opposed agency directed at a goal, with degrees of success—and omits the martial measurements that do not transfer. One may still prefer a definition of argument that takes evidence and conclusion as essential and treats the war framing as optional. That preference is a dispute about which characteristic explains the most, which is exactly Rand’s test for an epistemological essence. It is not a dispute about whether the metaphorical hierarchy is formed by quantitative similarity. It is.

Mother

If mother is defined only by giving birth, the other cases become analogical. If it is a radial category, those cases are licensed extensions from an idealized model. The Randian reconstruction suggested above treats the word as holding a relational structure—originating and sustaining a child—across different measures of biological, legal, and daily involvement. Adoptive and birth mothers then stand to one another as Rand says similars stand: same characteristic, different measure. What remains recalcitrant is any extension that keeps the word while dropping the relational structure entirely. If such extensions exist and still count as members in good standing, the object-level category outruns a single denominator. The theoretical concept radial category does not. It keeps the structure “center plus extensions” and omits the particular kinship system.

Women, fire, and dangerous things

The classifier balan is the hardest case for the identification, and it should be left hard. If women, fire, and dangerous things share no commensurate attribute, the grouping is not a Randian concept. It may still be a motivated chain, a cultural instrument, a grammatical convenience. Lakoff’s concept of that chain, however, is Randian in form. He isolates a distinguishing structure—a noun class organized by gender, danger, and mythic association—and he omits the particular nouns. The example shows the limit of the object-level claim. It does not show that the method by which the example is theorized lies outside concept-formation by quantitative similarity. It shows the opposite. Without that method there would be no example, only a list.

VII. Conclusion

Lakoff and Johnson share with Rand the thought that concepts exist to condense untamed experience into manageable elements. The present essay has argued for a stronger continuity. The methods they use to characterize hierarchy in language, and the use they make of metaphor to construct and extend that hierarchy, have essential characteristics. Those characteristics can be stated in Rand’s terms. A conceptual metaphor is a class of mappings sharing a cross-domain projection of inferential structure, with particular sources, targets, and lexical forms omitted. A basic-level category is a taxonomic cut sharing gestalt and motor structure, with particular members omitted. A prototype is a central case relative to which other members differ in degree. A radial category is a network with a center and licensed extensions. Hierarchy itself is degree along scales of inclusiveness, typicality, and schematicity. Metaphor builds hierarchy by retaining a common relational denominator and omitting the measurements that do not transfer.

That is Rand’s method applied to linguistic and cognitive structure. It does not erase every disagreement. Chained categories without a single denominator, networks with several centers, and Lakoff and Johnson’s official hostility to definition by essentials remain. Those are disputes about how far folk classification follows the method, not about whether the method is in use when the theorists isolate their own kinds. A theory of non-essentialist categories is still a theory of a kind of category. To have that kind in mind as one unit, to give it a word, and to distinguish it from classical categories, basic-level categories, and mere lists, is to do what Rand says a mind must do if it is to think at all.

The names continue to mislead. Lakoff and Johnson’s “objectivism” is not Rand’s Objectivism.[34] One may reject a doctrine of disembodied symbols mapped onto pre-sorted kinds and still hold that concept-formation isolates a commensurate characteristic and omits its particular measure.[35] Once that distinction is kept, the comparison yields a precise result. Rand explains how a mind holds open-ended classes as one unit. Lakoff and Johnson explain how a mind holds unbounded and abstract experience by giving it hierarchical and metaphorical form. The second explanation, when it is made into a theory, is an instance of the first.

Works Cited

Dixon, R. M. W. The Dyirbal Language of North Queensland. Cambridge: Cambridge University Press, 1972.

Lakoff, George. Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago: University of Chicago Press, 1987.

Lakoff, George, and Mark Johnson. Metaphors We Live By. Chicago: University of Chicago Press, 1980. With a new afterword, 2003.

Rand, Ayn. Introduction to Objectivist Epistemology. Expanded 2nd ed. Edited by Harry Binswanger and Leonard Peikoff. New York: Meridian, 1990.

Rand, Ayn. “The Objectivist Ethics.” In The Virtue of Selfishness, 13–35. New York: Signet, 1964.

Rosch, Eleanor. “Principles of Categorization.” In Cognition and Categorization, edited by Eleanor Rosch and Barbara B. Lloyd, 27–48. Hillsdale, NJ: Lawrence Erlbaum, 1978.

Wittgenstein, Ludwig. Philosophical Investigations. Translated by G. E. M. Anscombe. 3rd ed. Oxford: Blackwell, 1967.

[1]Ayn Rand, Introduction to Objectivist Epistemology, expanded 2nd ed., ed. Harry Binswanger and Leonard Peikoff (New York: Meridian, 1990). First serialized in The Objectivist (July 1966–February 1967).

[2]George Lakoff and Mark Johnson, Metaphors We Live By (Chicago: University of Chicago Press, 1980; with a new afterword, 2003); George Lakoff, Women, Fire, and Dangerous Things: What Categories Reveal about the Mind (Chicago: University of Chicago Press, 1987).

[3]Rand, Introduction to Objectivist Epistemology, 13.

[4]Ibid., 13–15. The grasp of similarity, Rand holds, requires contrast with a third item that differs along the same scale of measurement, the “conceptual common denominator.”

[5]Ibid., 10.

[6]Ibid., 10–11, 40.

[7]Ibid., 40–54. Essence is epistemological rather than metaphysical: the characteristic that, within the present context of knowledge, explains the most and distinguishes the units from everything else so far identified.

[8]Ibid., 62–65.

[9]Ibid., 65–66.

[10]Ibid., 69.

[11]Lakoff and Johnson, Metaphors We Live By, 3, 5.

[12]Ibid., 25.

[13]Ibid., 10–13, 163.

[14]Lakoff, Women, Fire, and Dangerous Things, 12–57, drawing on Eleanor Rosch, “Principles of Categorization,” in Cognition and Categorization, ed. Eleanor Rosch and Barbara B. Lloyd (Hillsdale, NJ: Lawrence Erlbaum, 1978), 27–48.

[15]Lakoff, Women, Fire, and Dangerous Things, 68–76, 91–114.

[16]Ibid., 92–104; R. M. W. Dixon, The Dyirbal Language of North Queensland (Cambridge: Cambridge University Press, 1972).

[17]Lakoff and Johnson, Metaphors We Live By, 185–222; Lakoff, Women, Fire, and Dangerous Things, 157–195.

[18]Rand, Introduction to Objectivist Epistemology, 52–54.

[19]This is a version of a familiar objection to strong anti-essentialism: the descriptive vocabulary in which one denies essences must itself pick out kinds. The point here is narrower. It is not that Lakoff and Johnson contradict themselves by speaking. It is that the kinds they isolate—metaphor, basic level, radial category, idealized cognitive model—are formed by the same operations Rand takes to be concept-formation as such.

[20]Lakoff and Johnson, Metaphors We Live By, 5: “The essence of metaphor is understanding and experiencing one kind of thing in terms of another.” The word “essence” is theirs.

[21]Ibid., 52–55, on the partial nature of metaphorical structuring; 147–155, on the creation of similarity.

[22]Rand, Introduction to Objectivist Epistemology, 11–15. What metaphor omits is not always a magnitude on a single scale, but it is particularity along dimensions the mapping does not carry. That is measurement-omission generalized from extensive magnitudes to relational structure.

[23]Lakoff, Women, Fire, and Dangerous Things, 12–13, 31–40. Rosch’s typicality ratings are already quantitative: subjects assign degrees of membership or goodness-of-example.

[24]Ibid., 283–317, on the Spatialization of Form Hypothesis: category structure itself is understood in terms of image schemas such as CONTAINER, PART-WHOLE, CENTER-PERIPHERY, and UP-DOWN.

[25]Lakoff and Johnson, Metaphors We Live By, 14–21, on MORE IS UP and other orientational metaphors grounded in bodily correlation.

[26]Rand, Introduction to Objectivist Epistemology, 29–38, on abstraction from abstractions and the hierarchical structure of knowledge.

[27]Ludwig Wittgenstein, Philosophical Investigations, trans. G. E. M. Anscombe, 3rd ed. (Oxford: Blackwell, 1967), §§66–67; Lakoff, Women, Fire, and Dangerous Things, 12–22, 91–104.

[28]Lakoff, Women, Fire, and Dangerous Things, 74–84.

[29]Rand, Introduction to Objectivist Epistemology, 72–74.

[30]Lakoff and Johnson, Metaphors We Live By, 115–125, reject definition as a list of necessary and sufficient conditions and treat understanding as organized by prototypes and gestalts.

[31]Rand, Introduction to Objectivist Epistemology, 65–69.

[32]Lakoff and Johnson, Metaphors We Live By, 25–32.

[33]Ibid., 4–6, 10–13.

[34]On the distinction between Lakoff’s “objectivism” and Rand’s Objectivism, see Lakoff and Johnson, Metaphors We Live By, 185–194; Rand, Introduction to Objectivist Epistemology, 1–4, 55–61; Ayn Rand, “The Objectivist Ethics,” in The Virtue of Selfishness (New York: Signet, 1964), 13–35.

[35]A related attempt to distinguish the object of awareness from the means of awareness, without collapsing Lakoff’s target into Rand’s system, is “Beyond the Introduction to Objectivist Epistemology,” Every Waking Moment (2026). The present essay goes further in one respect: it treats Lakoff and Johnson’s theoretical kinds as themselves products of measurement-omission.