Education Before Training
Liberal formation, professional competence, and the institutional purpose of the university
Why a free society should distinguish the formation of judgment from occupational preparation—and why mass credentialism, ideological closure, and artificial intelligence make that distinction urgent
Abstract
Contemporary educational policy frequently treats education, training, credentialing, employability, and social mobility as interchangeable objectives. This paper argues that the resulting conceptual compression has damaged both universities and vocational institutions because it assigns heterogeneous purposes to a single institutional form. Drawing on the classical tradition of liberal education, modern philosophy of education, the economics and sociology of credentialing, and comparative research on vocational formation, I distinguish education as the formation of judgment from training as the acquisition of task- and practice-specific competence. The distinction is analytical rather than hierarchical: advanced professional practice requires both, while high-quality apprenticeship can contain significant educational content. The institutional implication is nevertheless substantial. A defensible system should provide a demanding common education for all, preserve selective advanced academic education as a distinctive practice, locate professional formation on top of substantial general intellectual formation, and restore apprenticeships and technical routes as high-status alternatives rather than failed versions of university. The argument also addresses the strongest objections: that selective academic education is inherently elitist, that liberal education can become culturally exclusionary, that vocational routes can reproduce class stratification, and that criticism of ideological conformity can become a partisan project. The conclusion is that the relevant normative standard is not universal participation in one institutional pathway but universal access to intellectual foundations combined with plural, permeable routes to excellence. Artificial intelligence strengthens rather than weakens this case because the automation of routine intellectual procedures increases the relative value of judgment, explanation, methodological discrimination, and the capacity to challenge premises.
Keywords: liberal education; vocational education; apprenticeship; credentialism; human capital; academic freedom; professional education; artificial intelligence; university purpose
Introduction: the category error at the centre of modern education
Modern educational systems are asked to produce citizens, workers, professionals, researchers, social mobility, economic growth, personal flourishing, and credentials that permit employers to sort applicants. The problem is not that any one of these aims is illegitimate; the problem is that institutions increasingly behave as though they were identical. Labaree (1997) captured the resulting tension in his distinction among democratic equality, social efficiency, and social mobility, showing that educational systems are pulled simultaneously toward public civic purposes, labour-market purposes, and private positional advantage. Biesta (2010) reaches a parallel conclusion from philosophy of education by distinguishing qualification, socialization, and subjectification, while Brighouse et al. (2018) identify a wider set of educational goods that includes productivity, autonomy, democratic competence, relationships, equal standing, and personal fulfilment. These literatures differ in method and normative commitments, but they converge on a central point: education is not one thing, and institutions become incoherent when a single metric or function is allowed to stand in for the whole. The present argument begins from that pluralism but adds a sharper institutional distinction between education and training because the failure to preserve that distinction has encouraged both the vocationalization of universities and the denigration of vocational routes.
The central claim is that education concerns the formation of judgment through initiation into bodies of knowledge and disciplined modes of inquiry, whereas training concerns the acquisition of reliable competence within a specified practice. The distinction does not imply that education is intrinsically superior to training, because societies depend on forms of expertise that are learned through practice, repetition, correction, and tacit knowledge as much as through abstract study. Nor does it imply that academic disciplines are non-practical, since mathematics, law, engineering, medicine, and the sciences necessarily combine conceptual understanding with procedural competence. The distinction instead identifies different centres of gravity: education asks what is true, why a method is warranted, which assumptions structure an inquiry, and when established procedures should be revised; training asks how a task is to be performed competently under known standards and constraints. Professional formation occupies the overlap because professional judgment requires both principled understanding and the capacity to act competently in situations that rarely reproduce textbook conditions exactly. Once the categories are distinguished, a more coherent institutional architecture becomes possible.
This distinction also explains why the standard defence of mass higher education is incomplete. If every occupational pathway is translated into a degree pathway, the university must simultaneously function as a site of intellectual formation, a provider of occupational competence, and a labour-market screening device. Becker’s (1993) theory of human capital explains why individuals and states rationally invest in education when knowledge and skills raise productivity, while Spence’s (1973) signalling model explains why credentials can retain private value even when their educational content is not the sole source of that value. Collins (2019) and Labaree (1997) show the institutional consequence: as credentials become necessary for positional competition, qualification requirements can escalate independently of the substantive complexity of work. The result is not universal enlightenment but credential inflation, academic drift, and the conversion of educational institutions into sorting mechanisms. A system built around differentiated but permeable routes offers a way to preserve the goods of advanced academic study without forcing every form of occupational preparation into the same institutional container.
The argument therefore rejects two symmetrical errors. The first is the economistic reduction of education to employability, in which the value of a curriculum is inferred from wage returns or employer demand. The second is the romantic reduction of education to detached contemplation, in which practical competence is treated as intellectually inferior. Dewey (1916) opposed precisely this dualism, arguing that divisions between supposedly liberal and vocational learning become destructive when they separate reflection from intelligent action. The institutional goal should consequently be neither a university for everyone nor vocational tracking for those judged academically weaker, but a system in which everyone receives serious general education and then enters pathways whose purposes are explicit, demanding, and revisable. That model is academically selective without being socially hereditary, vocational without being anti-intellectual, and liberal without being ideologically vacant.
Figure 1. Analytical distinction among education, training, and professional formation. Education is organised primarily around knowledge, reasons, judgment, and the evaluation of ends; training is organised primarily around reliable performance within practices; professional formation combines both and adds responsibility under conditions of uncertainty. The distinction concerns dominant purpose rather than mutually exclusive institutional categories.
Education as the formation of judgment
The philosophical case for distinguishing education from training is strongest when education is defined neither as generic “critical thinking” nor as possession of cultural capital, but as entry into reason-governed practices of inquiry. Peters (1966) argued that education involves initiation into worthwhile modes of thought and conduct whose standards are internal to the activities themselves, a formulation that resists both behaviourist reduction and purely instrumental justification. The contemporary Stanford Encyclopedia of Philosophy similarly describes a central tradition in which education develops the ability and disposition to seek reasons, evaluate them, and allow judgment to be guided by that evaluation (Curren, 2025). This matters because “thinking skills” cannot be separated from substantive knowledge: one cannot reason historically without historical evidence and chronology, evaluate a statistical claim without mathematics and probability, or judge a legal argument without understanding institutions, doctrine, and language. Education therefore requires both content and the intellectual practices through which content is made intelligible. Its defining achievement is not information possession but the capacity to move responsibly among evidence, inference, explanation, and judgment.
Plato’s Republic provides an early and influential model of this formative conception even though its political architecture cannot serve as a template for a modern liberal society. In Book VII, education is represented as a reorientation of the soul rather than the insertion of knowledge into an otherwise empty mind, and the sequence from mathematics to dialectic is designed to cultivate the capacity to move beyond appearances toward structured reasons (Plato, 1992). The political hierarchy, censorship, and exclusion built into Plato’s wider scheme should be rejected, but the educational insight survives that rejection: advanced understanding requires cultivated intellectual habits rather than the accumulation of immediately useful techniques. Aristotle develops a related distinction in Book VIII of the Politics, where he differentiates learning pursued merely for utility from education appropriate to people capable of free judgment (Aristotle, 2013). The ancient social context made freedom a privilege rather than a universal status, yet modern liberalism can universalise the status while retaining the conceptual distinction. The educational task becomes the cultivation of capacities required for self-direction, not the preservation of an inherited social order.
The tradition of liberal education becomes institutionally important because “liberal” here refers to freedom rather than to a contemporary partisan alignment. Oakeshott (1989) presents education as entry into a conversation composed of different modes of human understanding, while Newman (1852/1996) defends liberal knowledge against the claim that university study must justify itself through immediate utility. Neither argument requires hostility to work, commerce, or professional competence. Their point is that a university ceases to perform a distinctive educational function when every intellectual activity is valued primarily as an input into an occupational outcome. Mill stated the sequence with unusual clarity in his 1867 St Andrews address: people are formed as human beings before they become lawyers or physicians (Mill, 1867). Read together, these arguments support an institutional principle rather than an antiquarian curriculum: professional specialization should operate upon minds already capable of reasoning beyond the boundaries of the profession.
The most serious contemporary version of this claim is not a defence of canonical knowledge for its own sake but a theory of educational goods. Brighouse et al. (2018) argue that educational institutions should develop capacities for economic productivity, personal autonomy, democratic competence, healthy relationships, treating others as equals, and personal fulfilment. Their framework is important because it blocks a simplistic opposition between liberal and useful education: economic productivity is a legitimate educational good, but it is one among several goods and can conflict with others. Biesta’s (2010) tripartite account makes a similar point in different language by distinguishing qualification from socialization and subjectification, thereby resisting the idea that learning to “do something” exhausts educational purpose. The consequence for curriculum is that intellectual formation cannot be measured exclusively by occupational placement, salary, or task competence. A curriculum should instead be judged partly by whether it enlarges the learner’s capacity to understand, evaluate, and act in relation to a world that cannot be reduced to employment.
This formulation also clarifies why mathematics and science belong within liberal education rather than standing outside it as merely technical subjects. Mathematics disciplines abstraction, proof, precision, counterexample, and the explicit statement of assumptions; science disciplines inference from observation, measurement, causal reasoning, uncertainty, and model revision. History, literature, philosophy, and the social sciences cultivate different but equally structured forms of reasoning concerning context, interpretation, institutions, meaning, incentives, and normativity. The educational point is therefore methodological plurality under standards of justification, not the elevation of humanities over STEM. A genuinely educated person need not master every discipline, but should understand that different questions require different evidential and inferential structures. That capacity to choose and evaluate forms of reasoning is precisely what narrow occupational instruction cannot reliably provide on its own.
Historical differentiation: liberal arts and the higher faculties
The medieval university does not supply a ready-made model for contemporary reform, but its institutional differentiation is historically relevant because it demonstrates that general intellectual formation and advanced professional study need not be collapsed into one undifferentiated credential. Universities differed markedly across Europe: Bologna developed around law, Paris around arts and theology, institutional structures changed over time, and many students did not follow a uniform sequence from one faculty to another. Haskins (1923) and de Ridder-Symoens (1992) nevertheless document the durable distinction between the Faculty of Arts and the higher faculties of theology, law, and medicine. The arts curriculum, conventionally associated with the trivium of grammar, rhetoric, and logic and the quadrivium of arithmetic, geometry, music, and astronomy, supplied forms of linguistic, logical, mathematical, and natural-philosophical preparation that could precede more specialised study. The point should not be romanticised into the false claim that the medieval university operated like a modern American system of undergraduate education followed by graduate professional school. Its importance is conceptual: institutional design could recognise that capacities of reasoning and general intellectual formation were not identical to the specialised knowledge of a profession.
That historical distinction also corrects the modern tendency to identify liberal education with the humanities alone. The medieval liberal arts placed logic and mathematics at the centre of intellectual formation, and later university curricula expanded through encounters with Aristotelian natural philosophy and other bodies of knowledge. A contemporary reconstruction should therefore resist both the vocationalisation of the university and the sentimentalisation of liberal education as an exclusively literary project. The relevant inheritance is a plural conception of reason in which language, argument, number, demonstration, interpretation, and inquiry into nature are different but mutually informing intellectual practices. Such a curriculum would not reproduce medieval content or authority structures, but it would preserve the idea that professional specialisation is strengthened when students have already encountered several disciplined ways of making and testing claims. Historical differentiation thus supports the broader thesis of this paper without requiring historical nostalgia: the educational system can distinguish formation from specialisation while allowing the two to interact.
The university as an institution of intellectual freedom
If education concerns judgment rather than mere transmission, then the university must preserve conditions under which conclusions are not institutionally predetermined. Humboldt’s account of the university is useful here because it treats advanced study as engagement with knowledge that remains unfinished rather than as the delivery of settled propositions (von Humboldt, 1810/1970). The teacher therefore does not merely transfer a stock of information; teacher and student participate in inquiry whose standards may expose both to correction. This principle is not an argument for epistemic relativism, since inquiry requires standards precisely because some claims are better supported than others. It is instead an argument against institutional closure: an educational institution loses its distinctive function when students are rewarded for reproducing authorised conclusions rather than for demonstrating command of evidence and argument. Academic freedom is therefore not an ornamental privilege of professors but part of the epistemic infrastructure of education.
This point is especially important when discussing critical theory, postcolonial theory, feminist theory, critical race theory, Marxism, conservative thought, liberalism, nationalism, religious thought, or libertarianism. A serious university should not remove difficult or politically charged traditions merely because they offend a constituency, since exclusion by ideological veto would reproduce the same intellectual pathology from the opposite direction. The relevant distinction is between teaching a framework as an object of inquiry and using the framework as an unchallengeable moral grammar through which all inquiry must proceed. Duarte et al. (2015) argue that ideological homogeneity can create risks of confirmation bias and narrow the range of questions and assumptions considered within a research community, while Whittington (2021) makes the broader institutional case that ideological diversity can improve the knowledge-producing function of universities. Neither argument proves that every department must contain a predetermined partisan distribution, and neither establishes that minority viewpoints are epistemically superior. They do, however, support the weaker and more defensible claim that systematic absence of serious counterargument can impair error correction.
The academic standard should therefore be adversarial symmetry rather than ideological balance by quota. A course on Marx should present Marx strongly enough that students can understand why the theory attracted serious adherents, but it should also expose the framework to its strongest historical, economic, and philosophical objections. A course on Hayek should do the same, as should courses on feminist theory, postcolonial thought, religious ethics, conservative political theory, or any other contestable intellectual tradition. The assessment criterion should be whether the student can reconstruct arguments accurately, use evidence responsibly, distinguish empirical from normative claims, and answer objections rather than whether the student affirms the lecturer’s politics. This standard is compatible with strong moral commitments because moral seriousness does not require intellectual immunity. Indeed, education becomes more valuable when students learn to distinguish commitment from dogmatism and criticism from caricature.
The objection that a “canon” can itself become an instrument of ideological closure is therefore correct but not decisive. A liberal curriculum that simply replaces contemporary orthodoxy with an authorised list of “great books” would reproduce the same error in historical form. The defensible reason for reading Plato, Aristotle, Augustine, Aquinas, Hobbes, Locke, Hume, Smith, Kant, Mill, Marx, Nietzsche, Du Bois, Hayek, Arendt, Fanon, Rawls, and other major figures is not that the list constitutes a sacred inheritance but that these writers formulate arguments that have structured enduring problems. Their value lies partly in disagreement, because the student is compelled to inhabit incompatible conceptual systems and to identify where they conflict. A tradition becomes educational when it is treated as material for disciplined judgment rather than as a set of conclusions to be inherited intact. Liberal education is consequently better understood as structured exposure to rival forms of reasoning than as possession of a culturally fixed reading list.
From liberal formation to professional competence
The distinction between education and training becomes institutionally significant at the point of professional formation. Modern professions operate through bodies of specialised knowledge, formal rules, practical routines, institutional responsibilities, and standards of competence that cannot be acquired through general education alone. A physician must learn diagnosis, pharmacology, clinical procedure, and the management of risk; a lawyer must learn doctrine, procedure, evidence, drafting, and advocacy; an engineer must master modelling, design standards, materials, and failure modes. These are not lesser forms of learning merely because they are professionally directed. They are forms of disciplined competence whose social value depends on reliability. The mistake is to assume that competence eliminates the need for prior intellectual formation.
Professional judgment is necessary precisely because real cases are underdetermined by procedure. Schön (1983) described professional practice as involving “reflection-in-action” in situations where the practitioner must frame the problem, interpret incomplete feedback, and revise action while operating. This is obvious in medicine, where evidence applies to heterogeneous patients; in law, where rules interact with facts, precedent, language, and institutional purpose; and in engineering, where models simplify systems that fail in particular environments. General intellectual formation matters because it gives professionals resources for recognising when a familiar procedure is being applied outside its domain. Historical knowledge can reveal why an institution has a particular rule, philosophy can expose hidden normative assumptions, statistics can identify unsupported causal inference, and literature can improve the interpretation of language and human motive. The profession becomes safer when its members can reason outside the procedural shell of the profession.
Mill’s formulation that human beings precede lawyers and physicians should therefore be interpreted as a sequencing principle rather than a claim about one universal degree structure. In the United States, law and medicine are generally graduate-entry professions, while in Britain and many other systems students can enter professional study directly from secondary education. There is no historical or empirical basis for claiming that one bureaucratic sequence is always optimal. The stronger claim is functional: professional curricula should contain or build upon substantial general education in reasoning, writing, quantitative methods, history, institutions, and ethics. An integrated six-year programme may satisfy that requirement as effectively as a three-year liberal degree followed by a professional degree if its early stages genuinely cultivate general intellectual capacities. What matters is not the label attached to the credential but whether professional specialization crowds out the formation of judgment.
This is also where universities have often confused status with purpose. If every occupation seeking greater status seeks degree status, and every professional association seeks university validation, academic institutions become the default certification mechanism for increasingly heterogeneous forms of training. The process can be self-reinforcing because degree status may improve labour-market position even when the occupational task has not changed proportionately. Collins (2019) describes this as credential inflation, while Labaree (1997) shows how social mobility as a private educational goal can convert education into a positional good whose value depends partly on relative scarcity. The individual decision to acquire a credential may remain perfectly rational under these conditions. The systemic problem is that private rationality can produce collective escalation without corresponding gains in intellectual or occupational capability.
Human capital, signalling, and the institutional logic of credentialism
The economic case for investment in education is powerful and should not be caricatured. Becker (1993) formalised the idea that education and training can increase productive capacities and therefore generate returns to individuals and societies. This framework explains why occupationally relevant knowledge matters, why firms invest in training, and why states have strong reasons to finance education when social returns exceed private returns. The problem arises only when human-capital theory is treated as a complete philosophy of education rather than as a theory of one important class of benefits. Wage returns cannot determine whether historical understanding, democratic competence, aesthetic appreciation, or intellectual autonomy are valuable because those goods are not reducible to productivity. The human-capital framework should therefore inform educational policy without monopolising its normative vocabulary.
Spence’s (1973) signalling model identifies a second mechanism that complicates the relationship between credentials and productive learning. When employers cannot directly observe characteristics such as persistence, cognitive ability, reliability, or conformity to institutional expectations, educational credentials can function as signals even where the educational process itself is not the sole source of productivity. This does not imply that education teaches nothing; human-capital formation and signalling can coexist within the same credential. The institutional consequence nevertheless matters because the private return to a degree can remain high even if part of that return arises from sorting rather than from skill acquisition. Once employers use degrees as screens, individuals rationally acquire degrees to avoid exclusion, and employers can then raise requirements because the enlarged pool permits them to do so. Credential inflation is therefore not evidence that students are irrational; it is a coordination problem generated by individually rational responses to a positional market.
Labaree’s (1997) account provides the educational counterpart to this economic mechanism. When social mobility becomes dominant, education is experienced less as a public practice of formation and more as a private commodity that allocates access to desirable social positions. Under those conditions, the content of the credential can become secondary to its scarcity and prestige, which encourages institutions to compete on brand while students compete on credential level. Collins (2019) extends this argument sociologically by showing how professional closure and escalating educational requirements can reproduce stratification even under formal expansion of access. Caplan (2018) advances a stronger and more controversial version in which signalling constitutes a large share of educational returns; his quantitative conclusions are contestable, but the institutional mechanism does not depend on accepting his highest estimates. Even a moderate signalling component is enough to explain why an economy can experience both rising educational attainment and persistent employer complaints about capability. The policy implication is to reduce unnecessary reliance on academic credentials for occupational screening rather than to assume that still more degree participation will solve the problem created by degree inflation.
This diagnosis also clarifies why a defence of selective advanced academic education need not be a defence of inherited privilege. If academic education is a positional good, then limiting access by wealth would intensify inequality and reproduce class advantage. The defensible principle is instead open access to competition combined with demanding standards and extensive support for those whose background restricts preparation. Brighouse et al. (2018) emphasise that educational policy must consider both the distribution of educational goods and independent values such as autonomy and opportunity. A system can therefore be selective in the level of advanced academic work while remaining egalitarian in the conditions under which students can attempt it. Selectivity becomes unjust when it tracks inherited status rather than relevant preparation and capability; it is not unjust simply because different forms of advanced activity require different levels of performance.
Apprenticeship and the knowledge embedded in practice
A coherent educational system must also stop treating apprenticeship as failed university. Apprenticeship is a distinctive institution because it combines formal instruction with participation in a practice whose standards are embodied in real work, real constraints, and correction by experienced practitioners. Wolter and Ryan (2011) show that apprenticeship systems vary substantially across countries and sectors in their financing, employer incentives, skill content, and quality, which means “apprenticeship” should not be treated as a single uniform model. Busemeyer and Trampusch (2012) further demonstrate that successful collective skill-formation systems depend on institutional arrangements among firms, associations, labour organisations, and the state. The quality of apprenticeship therefore depends on governance as much as on pedagogy. Well-designed systems create portable competence; poorly designed systems can become cheap labour or narrow firm-specific instruction.
The epistemic case for apprenticeship rests partly on the limits of explicit knowledge. Many occupational capacities are acquired through repeated perception, correction, timing, and judgment in circumstances where the practitioner cannot reduce every competent action to a proposition. A machinist learns to recognise tolerances, a technician learns how equipment behaves under imperfect conditions, a chef learns timing and material variation, and an electrician learns how abstract principles interact with the actual constraints of a site. This tacit dimension does not make vocational knowledge irrational; it means that competence is partly embodied in practice. Universities sometimes underestimate this because their dominant forms of assessment reward explicit verbal or symbolic representation. Treating practical knowledge as intellectually inferior merely because it is not best demonstrated in an essay is therefore a category error.
The strongest objection to vocational education is not that it lacks value but that it can become too narrow. Hanushek et al. (2017) find evidence of a life-cycle trade-off in which vocational education can improve early labour-market transition while general education may support adaptability later as technologies and occupational structures change. That finding is especially important in an era of rapid automation because task-specific competence can depreciate quickly when production systems change. The appropriate response is not to replace vocational education with universal university attendance, but to design vocational routes that contain substantial general education in mathematics, communication, science, digital reasoning, and relevant institutional knowledge. A technically trained worker who understands the principles underlying current procedures is better placed to adapt when those procedures change. The best apprenticeship system therefore combines training with education rather than choosing between them.
This argument also addresses the class-stratification objection that has historically accompanied differentiated pathways. Systems that direct affluent students toward broad academic formation while assigning disadvantaged students to narrow vocational tracks can reproduce inequality under the language of “fit.” That risk is real and should shape institutional design. Differentiation should therefore occur after a strong common foundation, entry to academic routes should remain open later in life, vocational qualifications should be portable, and bridges into higher education should be institutionalised rather than exceptional. The normative aim is differentiation without caste: different forms of excellence should be recognised without turning an early pathway decision into a permanent social identity. A plural system is just only if it remains permeable.
Figure 2. Institutional architecture for education, professional formation, and training. A common educational foundation precedes differentiated academic, professional, and vocational routes. The model is deliberately permeable: later university entry, professional conversion, and advanced technical progression prevent differentiated routes from hardening into inherited caste.
What should be universal, and what should be selective
The phrase “not everyone should go to university” is often heard as a claim about human worth because modern societies have attached status to academic credentials. That inference should be rejected. Equal dignity requires that every person be treated as an equal moral subject and given access to foundational educational goods; it does not require every person to undertake the same advanced activity. No society assumes that universal access to music requires universal conservatoire attendance, or that universal mathematical literacy requires universal study of measure theory. Advanced academic education can therefore be selective without denying anyone’s right to education, provided the selection is connected to the demands of the activity rather than to inherited social rank. The policy challenge is to separate selectivity of standard from exclusivity of opportunity.
The universal component should consequently be more intellectually demanding than many current systems provide. Every student should leave compulsory education able to read difficult prose, write coherent extended arguments, reason quantitatively, interpret probability, understand basic scientific inference, locate contemporary institutions historically, and distinguish empirical claims from normative ones. A common curriculum should also expose students to literature, arts, civics, economics, and competing political and moral traditions because citizenship in a free society requires more than occupational functionality. Brighouse et al. (2018) provide a useful justification by linking educational goods to autonomy, democratic competence, productive capability, and personal flourishing. This common foundation is the egalitarian core of the system because it gives everyone access to capacities needed for adult agency. Differentiation should occur after, not instead of, that foundation.
Advanced academic education should then be reserved for forms of study that genuinely require sustained abstraction, research, disciplinary immersion, or sophisticated theoretical work. The point is not to make universities smaller for symbolic reasons but to allow them to do something institutionally distinctive. If academic programmes must admit almost everyone and simultaneously function as universal employment passports, pressure grows to lower curricular differentiation, inflate credentials, and substitute generic employability content for disciplinary depth. A university that attempts to be school, professional body, corporate trainer, social-service institution, and labour-market sorting agency at once will predictably become unclear about its primary standards. Selective academic education is defensible when it protects depth while alternative routes receive comparable social respect and real economic opportunity. Selectivity without alternatives is exclusion; selectivity within a plural system can be institutional coherence.
The strongest egalitarian objection is that academic ability itself is socially produced, making selection appear neutral while reproducing differences in family resources, schooling, cultural capital, and expectations. That objection is compelling and rules out complacent meritocracy. Equal competition requires compensatory investment in early education, financial support, second-chance entry, transparent admissions, and routes that recognise late development rather than treating performance at eighteen as destiny. It also requires universities to distinguish relevant academic preparation from social signals that merely correlate with privilege. The answer to unequal preparation is not to abolish standards, because doing so can conceal rather than eliminate inequality. The more defensible approach is to invest heavily in the conditions under which students can meet demanding standards and to keep routes open for those who develop later.
The problem with ideological closure in higher education
A university committed to the formation of judgment must also examine how its own curricular and administrative practices can undermine that purpose. The relevant problem is not the existence of “critical studies” as a category, because critical traditions have generated substantial bodies of scholarship and often expose assumptions that conventional approaches overlook. The problem begins when a framework acquires institutional immunity from the ordinary demands of inquiry, so that disagreement is treated as evidence of moral deficiency rather than as an intellectual position to be tested. Biesta’s (2010) criticism of the “learnification” of education is relevant here because educational language can obscure questions of purpose and authority; one can likewise speak constantly of criticality while narrowing the range of permissible criticism. A curriculum does not become emancipatory merely by adopting the vocabulary of emancipation. The educational test is whether students are equipped and permitted to question the framework itself.
The epistemic standard should be consistent across political traditions. Marxist analysis should be tested against economic history, institutional evidence, and rival theories of incentives and coordination; market liberalism should be tested against evidence of market failure, power asymmetry, and distributional consequences; feminist and postcolonial theories should be studied in their strongest forms while remaining open to internal and external criticism; conservative and religious traditions should likewise be examined rather than insulated by deference to tradition. Duarte et al. (2015) show why homogeneous intellectual environments can create blind spots, while Whittington (2021) links ideological diversity to the broader knowledge mission of universities. The inference is not that truth lies at the midpoint between political camps, since evidence may decisively favour one side of a dispute. The inference is that institutions improve their chances of discovering error when serious counterarguments can be articulated without professional or pedagogical penalty.
This position must itself be constrained to avoid becoming a partisan programme disguised as academic freedom. Governments, donors, or administrators should not demand a predetermined ideological quota in scholarship, because such a requirement would politicise appointments and curricula from the opposite direction. Nor should “viewpoint diversity” become a licence to ignore disciplinary competence, methodological quality, or evidential standards. The appropriate safeguard is procedural: assessment criteria should be content-neutral where possible, controversial positions should be evaluated for argument and evidence, and scholarly disagreement should not be converted automatically into misconduct. A university should be capable of hosting asymmetrical conclusions when the evidence is asymmetrical while remaining symmetrical in the right to challenge those conclusions. Academic freedom protects inquiry, not every claim from criticism.
This is the point at which liberal education and freedom become inseparable. The educated person should become more, not less, capable of recognising the assumptions embedded in institutions, including the university itself. That capacity is politically inconvenient because trained compliance is easier to administer than independent judgment. A person may be excellently trained to implement a policy while lacking the conceptual resources to ask whether the policy’s objective is coherent, lawful, ethical, or empirically supported. Education creates the possibility of that second-order question. A free society therefore has reason to value education even when the immediate economic return is difficult to measure, because citizens and professionals must sometimes evaluate the systems within which they operate rather than merely perform efficiently inside them.
Artificial intelligence and the return of judgment
Artificial intelligence makes the education-training distinction more urgent because it lowers the cost of many procedural intellectual tasks. Systems can already generate routine prose, code, summaries, translations, statistical scripts, images, search strategies, and first-pass analyses at speeds that make traditional measures of effort increasingly unreliable. The wrong educational response is to preserve old assignments by prohibiting tools simply because the assignment once required manual labour. The correct response is to identify which cognitive capacity the assignment was intended to reveal and redesign assessment around that capacity. If an AI system can produce a technically plausible answer, the student should be required to explain why the method is appropriate, identify assumptions, verify sources, compare alternatives, and defend the conclusion under questioning. Automation should move assessment upward toward judgment rather than downward toward surveillance.
This does not eliminate the need for foundational competence. A student who cannot perform elementary algebra cannot reliably audit an AI-generated derivation, a programmer who has never written code cannot effectively evaluate generated software, and a lawyer who lacks doctrinal knowledge cannot identify invented authority. The relationship between education and AI is therefore sequential rather than substitutive. Learners require enough direct mastery to understand the structure of a task before automation can safely remove some mechanical execution. Once that threshold is reached, however, requiring manual repetition merely because it is traditional becomes difficult to defend educationally. The relevant question is whether the student understands the operation, not whether the student personally executed every step.
AI also strengthens the case for broad intellectual formation because automated systems are exceptionally good at producing locally fluent answers while remaining vulnerable to errors of framing, evidence, domain transfer, and objective specification. A model can optimise the wrong target with great efficiency, reproduce a false premise, apply a statistically correct technique to a substantively meaningless quantity, or generate a legally plausible statement unsupported by authority. Detecting those failures requires the very capacities that narrow training tends to neglect: epistemology, causal reasoning, history, institutional understanding, interpretation, and normative judgment. The scarce human contribution therefore shifts from routine production toward problem formulation and error detection. Education becomes more valuable precisely because training is increasingly automatable.
The implication for universities is severe. Institutions that respond to AI primarily by increasing invigilation and preserving assessment formats designed for a pre-AI world may protect the appearance of traditional work while missing the educational opportunity. Oral defence, adversarial questioning, source verification, methodological comparison, laboratory performance, fieldwork, live problem-solving, and iterative critique can all test understanding more directly than unsupervised production of routine text. Such assessments are more expensive, which is one reason mass systems may resist them. Yet that cost reveals the institutional trade-off that credential expansion has obscured: serious education requires sustained attention, feedback, and judgment, not merely content delivery. If society wants genuinely educated graduates, it must fund educational practices rather than assume that scale and quality are costless complements.
An institutional settlement: education first, training well
A coherent system should begin by making its purposes explicit. Compulsory schooling should provide broad education aimed at literacy, numeracy, scientific understanding, historical consciousness, civic competence, interpretive capacity, and intellectual autonomy. Post-compulsory routes should then differentiate according to purpose without being arranged as a hierarchy of human worth. Academic institutions should specialise in disciplinary depth, research, and advanced intellectual formation; professional institutions should combine general formation with practice-specific responsibility; vocational institutions should cultivate portable occupational competence through high-quality apprenticeship and technical education. Labour markets should reduce unnecessary degree requirements where competence can be assessed more directly. The resulting system would contain more institutional variety and less credential uniformity.
Universities would benefit from such differentiation because they could stop promising that every programme is simultaneously an intellectual vocation and an occupational guarantee. Some disciplines are valuable principally because they extend knowledge; others combine intellectual and professional aims; still others are better delivered through practice-rich institutions. The question should be asked honestly rather than decided by prestige. A programme whose central purpose is occupational competence should be evaluated substantially by whether it produces competent practitioners, while a research-oriented academic programme should be evaluated by disciplinary mastery and capacity for inquiry. Blurring these criteria allows weak programmes to shelter behind whichever justification is convenient. Clear purpose makes accountability more demanding rather than less.
Apprenticeship would also benefit because it would cease to be defined by what it is not. Instead of presenting vocational routes as alternatives for students who “failed” to reach university, policy should treat them as institutions with their own standards of excellence and their own intellectual requirements. Busemeyer and Trampusch (2012) demonstrate that strong skill-formation systems depend on collective governance, employer participation, portable certification, and links between vocational and general education. These features permit apprenticeship to become a genuine route to mastery rather than a narrow contract with one employer. Permeability is essential: advanced technicians should be able to progress into engineering or management, experienced practitioners should be able to enter university later, and graduates should be able to retrain through vocational routes without status loss. A plural system becomes credible only when movement among routes is normal rather than exceptional.
Professional education would become more demanding under this settlement because it could no longer confuse credential completion with professional judgment. Law, medicine, engineering, accounting, teaching, and management all require forms of reasoning that exceed procedural compliance. Professional programmes should therefore make their intellectual prerequisites explicit and assess whether students can transfer principles to unfamiliar cases. This may lengthen some forms of preparation, but cost should be evaluated systemically rather than by comparing only programme duration. A society that reduces unnecessary degree inflation while investing more deeply in genuinely consequential professional formation may spend less on positional credentials even if some professional routes become more rigorous. The relevant efficiency question is the allocation of educational resources, not simply the number of years attached to one credential.
The strongest objections and the limits of the argument
The first objection is historical: appeals to liberal education often idealise institutions that were elitist, exclusionary, and socially narrow. That criticism is correct and materially changes what can be inherited from the tradition. Plato and Aristotle cannot provide the social architecture of a modern educational system, medieval universities cannot be treated as if they were meritocratic, and nineteenth-century liberal education frequently operated within institutions closed to women, the poor, religious minorities, or colonised populations. The defensible inheritance is therefore conceptual rather than institutional: the distinction between formation and utility, the importance of reasoning across domains, and the idea that advanced intellectual work requires freedom from immediate instrumental demands. Modern education must universalise access to foundational capacities while discarding the social exclusions of the institutions that first articulated them. A return to liberal education is legitimate only as reconstruction, not restoration.
The second objection is sociological: differentiated routes can reproduce class even when formal access is open. Families with greater resources can purchase preparation, navigate admissions, understand professional pathways, and cushion the risks of long academic study, while less advantaged students may choose shorter vocational routes for reasons of immediate income rather than preference. This objection means that “choice” cannot be treated as self-validating. Scholarships, maintenance support, high-quality public schooling, late-entry mechanisms, and institutional bridges are necessary if pathways are to reflect capability and preference rather than liquidity constraints. It also means vocational routes must lead to real progression and earnings rather than serving as warehouses for those excluded elsewhere. Differentiation is defensible only when the routes are substantively good.
The third objection is philosophical: the distinction between education and training can be overstated because practices themselves can educate. Dewey (1916) is especially important here, since he rejects the separation of reflective intelligence from practical activity and warns that vocational education becomes stunted when it prepares people only for a fixed industrial role. This criticism requires a refinement of the thesis. Education and training are not mutually exclusive categories of experience; they are analytical distinctions among purposes and forms of learning. An apprenticeship can be deeply educational when it teaches principles, judgment, history, and adaptation, while a university course can amount to little more than training when it teaches students to reproduce a formula or ideological script. The institutional argument therefore concerns dominant orientation, not a rigid binary.
The fourth objection is political: criticism of ideological closure can itself become a campaign to privilege one side of contested politics. That risk is substantial because academic-freedom language is easily instrumentalised by governments and movements seeking control over curricula. The appropriate response is to defend procedural norms consistently rather than to legislate acceptable conclusions. Faculty should be free to reach politically inconvenient findings; students should be free to defend unpopular positions when they meet disciplinary standards; and institutions should remain free to distinguish scholarly argument from propaganda or incompetence. No ideology is owed representation independent of intellectual quality, but no ideology should be immunised from criticism by administrative power. The principle is open inquiry constrained by evidence and method.
The fifth objection is economic: advanced general education before professional specialization can be expensive, and a system that asks some students to study longer may intensify inequality. This is a real trade-off rather than a rhetorical problem. The answer is not to pretend that high-quality formation is cheap but to compare it with the cost of the existing credential arms race, remedial professional education, weak degrees used as labour-market screens, and repeated reskilling caused by narrow preparation. Hanushek et al. (2017) show why adaptability has labour-market value across the life course, while human-capital theory reminds us that education can be an investment whose returns accrue over decades. Some longer pathways may therefore be efficient when they produce durable judgment, while many current pathways may be inefficient despite being shorter if their main function is signalling. Institutional design should be evaluated over the life cycle rather than by immediate throughput.
Conclusion
The distinction between education and training is not a semantic preference; it is a theory of institutional purpose. Education forms judgment by initiating people into bodies of knowledge, standards of inquiry, and practices of reasoning that allow them to understand and evaluate the world rather than merely operate within a predefined procedure. Training develops competence in practices whose standards are more determinate and whose success can often be demonstrated in performance. Professional formation requires both because responsible practitioners must act competently while recognising when rules, models, or routines cease to fit the case before them. Once this distinction is restored, many contemporary educational confusions become easier to diagnose. Universities have been asked to carry functions that belong partly to professional bodies, vocational institutions, and labour-market screening, while apprenticeship has been treated as a lower-status approximation of university rather than as a distinct institution of skill formation.
The resulting reform programme is demanding but coherent. Everyone should receive a substantial common education because literacy, numeracy, scientific understanding, historical knowledge, civic competence, and reasoned judgment are conditions of adult agency in a free society. Advanced academic education should then be selective in standard but open in access, with social and financial institutions designed to prevent selectivity from reproducing inherited caste. Professional formation should be built upon substantial general intellectual capacity, whether through prior degrees or genuinely integrated curricula. Apprenticeships and technical education should be expanded, governed for quality, and connected to later academic and professional routes. Employers should reduce reliance on academic credentials where competence can be assessed directly, weakening the signalling mechanisms that drive credential inflation.
The university’s internal culture matters equally. Liberal education cannot survive if “critical” means only criticism permitted by an authorised theory, just as it cannot survive if governments or administrators impose a counter-orthodoxy in the name of balance. The standard must be intellectual: strong arguments, accurate evidence, methodological competence, and genuine exposure to contestation. Students should leave university more capable of identifying why an argument fails even when they want its conclusion to be true, and more capable of recognising evidence even when they dislike the conclusion it supports. That is a demanding account of freedom because it denies every political faction the right to convert education into affirmation. It is also the best defence of the university’s distinctive place in a democratic society.
Artificial intelligence finally makes the argument unavoidable. As routine intellectual procedures become cheaper, institutions that define education by the production of routine outputs will find their assessments and credentials increasingly detached from the capacities they claim to certify. The durable human contribution lies in framing problems, selecting methods, evaluating evidence, detecting hidden assumptions, integrating incompatible forms of knowledge, and exercising judgment under uncertainty. Those capacities are not mystical and they are not produced by slogans about “critical thinking”; they require demanding education in substantive disciplines and repeated exposure to serious disagreement. Training will remain indispensable because societies require competent action. But when technology can perform more of the procedure, the value of understanding why the procedure exists—and when to reject it—rises rather than falls.
A mature educational system should therefore stop measuring equality by the proportion of people who pass through one institutional pathway. Equality of dignity is compatible with plurality of excellence, and plurality is more credible when routes remain permeable and standards remain real. The purpose of education is not to turn every citizen into an academic, nor to turn universities into universal job-training centres. It is to give people the intellectual foundations required for freedom and then to allow different forms of advanced formation to develop according to their purposes. Educate broadly, educate rigorously, train professionally, and stop pretending that these activities become equal by being called the same thing.
References
Aristotle. (2013). Aristotle’s politics (C. Lord, Trans.; 2nd ed.). University of Chicago Press. (Original work published ca. 350 B.C.E.)
Becker, G. S. (1993). Human capital: A theoretical and empirical analysis, with special reference to education (3rd ed.). University of Chicago Press. https://doi.org/10.7208/chicago/9780226041223
Biesta, G. J. J. (2010). Good education in an age of measurement: Ethics, politics, democracy. Paradigm Publishers.
Brighouse, H., Ladd, H. F., Loeb, S., & Swift, A. (2018). Educational goods: Values, evidence, and decision-making. University of Chicago Press. https://doi.org/10.7208/chicago/9780226514208.001.0001
Busemeyer, M. R., & Trampusch, C. (Eds.). (2012). The political economy of collective skill formation. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199599431.001.0001
Caplan, B. (2018). The case against education: Why the education system is a waste of time and money. Princeton University Press. https://doi.org/10.23943/9781400889327
Collins, R. (2019). The credential society: An historical sociology of education and stratification (Legacy ed.). Columbia University Press. (Original work published 1979)
Curren, R. (2025). Philosophy of education. In E. N. Zalta & U. Nodelman (Eds.), The Stanford encyclopedia of philosophy. Stanford University. https://plato.stanford.edu/entries/education-philosophy/
Dewey, J. (1916). Democracy and education: An introduction to the philosophy of education. Macmillan.
Duarte, J. L., Crawford, J. T., Stern, C., Haidt, J., Jussim, L., & Tetlock, P. E. (2015). Political diversity will improve social psychological science. Behavioral and Brain Sciences, 38, e130. https://doi.org/10.1017/S0140525X14000430
Hanushek, E. A., Schwerdt, G., Woessmann, L., & Zhang, L. (2017). General education, vocational education, and labor-market outcomes over the lifecycle. Journal of Human Resources, 52(1), 48–87. https://doi.org/10.3368/jhr.52.1.0415-7074R
Haskins, C. H. (1923). The rise of universities. Henry Holt and Company.
de Ridder-Symoens, H. (Ed.). (1992). A history of the university in Europe: Vol. 1. Universities in the Middle Ages. Cambridge University Press.
Labaree, D. F. (1997). Public goods, private goods: The American struggle over educational goals. American Educational Research Journal, 34(1), 39–81. https://doi.org/10.3102/00028312034001039
Mill, J. S. (1867). Inaugural address delivered to the University of St. Andrews, February 1st 1867. Longmans, Green, Reader, and Dyer. https://en.wikisource.org/wiki/Page:Inaugural_address_delivered_to_the_University_of_St._Andrews,_Feb._1st_1867.djvu/8
Newman, J. H. (1996). The idea of a university (F. M. Turner, Ed.). Yale University Press. (Original work published 1852)
Oakeshott, M. (1989). The voice of liberal learning: Michael Oakeshott on education (T. Fuller, Ed.). Yale University Press. https://doi.org/10.2307/j.ctt1xp3tm1
Peters, R. S. (1966). Ethics and education. George Allen & Unwin.
Plato. (1992). Republic (G. M. A. Grube, Trans.; C. D. C. Reeve, Rev.). Hackett Publishing Company.
Schön, D. A. (1983). The reflective practitioner: How professionals think in action. Basic Books.
Spence, M. (1973). Job market signaling. The Quarterly Journal of Economics, 87(3), 355–374. https://doi.org/10.2307/1882010
von Humboldt, W. (1970). On the spirit and organisational framework of intellectual institutions in Berlin. Minerva, 8(2), 242–250. (Original work written 1810)
Whittington, K. E. (2021). The value of ideological diversity among university faculty. Social Philosophy and Policy, 37(2), 90–113. https://doi.org/10.1017/S0265052521000066
Wolter, S. C., & Ryan, P. (2011). Apprenticeship. In E. A. Hanushek, S. Machin, & L. Woessmann (Eds.), Handbook of the economics of education (Vol. 3, pp. 521–576). Elsevier. https://doi.org/10.1016/B978-0-444-53429-3.00011-9