Higher education measures what it can count with remarkable precision. Institutions track attendance hours, enrollment numbers, transfer rates, credit accumulation and degree completion down to the decimal. These figures are tangible and consequential. No one debates what counts as enrollment; no one disputes what qualifies as a transfer; no one questions what constitutes a degree; yet the central purpose those metrics are meant to serve—learning—receives far less clarity. What can learners do now that they could not do before? Higher education can answer questions about enrollment and completion to the decimal. This one, it mostly cannot.
This asymmetry undermines the very value of the degree. Participation is quantified. Progression is documented. Completion is officially recognized. Learning, by contrast, is often inferred from grades, and its definition varies across courses, programs and institutions. Colleges can specify exactly what constitutes admission or completion. Why is there less agreement about what counts as learning?
AI Makes the Ambiguity Harder to Ignore
For decades, that ambiguity may have seemed manageable, but it has become more consequential in the age of generative artificial intelligence. When students can produce polished essays, solve equations or summarize readings with minimal effort, the distinction between performance and competence grows fragile. If institutions cannot articulate what learned capability looks like, they risk certifying outputs without verifying demonstrated capability. Public trust in higher education rests not only on graduation rates but on confidence that graduates can do what their credentials imply. The issue is not whether learning occurs. It is whether institutions can specify and demonstrate what it is.
Higher Education Has No Single Definition of Learning
Part of the difficulty lies in the way higher education talks about learning. The academy draws from multiple intellectual traditions, each emphasizing different aspects of what it means to learn. That diversity has deepened our understanding of thinking and development. It has also produced definitional variation.
Cognitive psychology often describes learning as an internal process in which students build mental models and deepen understanding. These explanations are valuable, but they describe processes that institutions cannot directly observe, so they infer learning from performance indicators that may or may not capture underlying change. Constructivist traditions emphasize meaning making within context. Learning is something students actively construct, a perspective that foregrounds agency, yet when meaning is individualized, establishing shared criteria becomes more complex. Behavioral traditions take a more direct route. They define learning as a change in observable performance resulting from experience or instruction. This approach focuses on what is demonstrable and verifiable. It may appear narrow to some, but it aligns naturally with systems that require evidence. It is also the only one of the three that can answer the question directly: What can learners do now that they could not do before?
These differences are not abstract debates. They shape course design and the way institutions attempt to determine whether students have learned. When no shared criteria exist for what learning must look like in practice, institutions rely on local interpretation. Statements about student learning may be carefully written, yet what students must be able to demonstrate is redefined one committee at a time. The issue is not which theory is philosophically correct. The issue is which definition can anchor accountability. Without agreement about what learning must look like in observable terms, assessment of student learning risks becoming procedural rather than evidentiary.
Teaching Strategies Are Not Evidence of Learning
This tension extends into contemporary pedagogical reform. Active learning, student-centered teaching and project-based instruction have reshaped classroom practice. They describe how instruction should unfold and how students should participate. What they define less clearly is what students should be able to demonstrate at the end.
Active learning prioritizes interaction, a valuable condition for growth, but participation alone does not specify competence. A student may contribute to a group project without independently demonstrating the skill the activity was designed to develop. Student-centered instruction promotes autonomy and individualized pathways, yet institutions still confront a basic question: What should all students, regardless of path, be able to do upon completion? Experiential and culturally responsive approaches enrich context and relevance, but when emphasis rests primarily on process, the explicit articulation of competence can recede into the background. The point is not to question these approaches. They are strategies, and strategies describe how learning may occur. Accountability requires clarity about what learning looks like when it has occurred.
When that clarity is missing, institutions often evaluate instruction rather than what students can demonstrably do. Faculty development focuses on method. Peer review highlights classroom activity. Documentation of student learning assessment confirms that activities took place. Whether students can consistently demonstrate transferable capability becomes harder to determine.
Grades Cannot Carry the Weight of Accountability
The same looseness extends into accountability systems. Accrediting bodies require institutions to publish student learning outcomes. Programs collect evidence of student learning. Reports are generated. However, in many cases, the emphasis rests on documenting that assessment of student learning occurred rather than on establishing shared criteria for what learning must look like across contexts. Grades have become a convenient mask for a lack of measurable skill. They compress performance, effort and participation into a single symbol. Students interpret them as evidence of learning, and institutions rely on them to signal completion, but grades do not define learning. They represent performance within a course structure. They do not reliably indicate transferable capability beyond it.
To be clear, assessment of student learning is not absent from higher education. Faculty ask follow-up questions, administer exams and evaluate student work. The issue is not whether student learning is evaluated. It is whether those evaluations align with the competencies institutions claim to develop. Any competency a program identifies as an intended learning outcome for its graduates must ultimately be evidenced through performance. Public speaking requires speaking. Engineering design requires designing. Clinical competence requires performing. Even capacities such as analysis or evaluation are inferred from what students can demonstrably do.
Ambiguity Has Consequences for Students and Institutions
When learning lacks shared and explicit criteria, the consequences ripple outward. Students rely on the clearest signals available to them. In the absence of a stable definition of competence, grades, effort and instructor approval become the primary indicators of progress. Students complete courses and accumulate credits without a clear articulation of what they can now do that they could not do before. Faculty operate within the same ambiguity. Asked to assess student learning without shared criteria, they often rely on constructs such as engagement or comprehension. These describe elements of classroom experience, but they do not consistently translate into transferable demonstrations of capability. At the institutional level, the effects accumulate. Systems track enrollment, persistence and completion with precision, while the definition of learning varies across programs and disciplines. Institutions can report that student learning has been assessed. Whether learning is consistently defined and verified remains uncertain. In this context, definitional clarity is not an academic preference. It is a structural requirement.
Make Demonstrable Learning the Anchor
If learning is to anchor accountability, it must be defined in terms that are observable and shared. One workable definition is straightforward: Learning is a change in observable performance resulting from experience or instruction that persists over time. That change is always the acquisition of a skill. Applying a concept in a new context is a skill. Refining a process with greater precision is a skill. What matters is not whether a student reports understanding but whether the student can demonstrate capability that was not previously present. This definition does not exclude reflection or intellectual growth. It requires evidence of them. Writing an argument, formulating a hypothesis, conducting an analysis or designing a solution are not superficial acts. They reveal reasoning and judgment in action.
Defining learning in demonstrable terms strengthens alignment across systems. It clarifies expectations for students, provides faculty with explicit targets for instruction and grounds credentials in observable capability rather than symbolic indicators. Higher education does not lack commitment to learning. What it lacks is a shared operational definition that can anchor accountability. An institution that tracks enrollment to the decimal but cannot say what its graduates can do now that they could not do before has measured everything except its purpose. Naming learning in demonstrable terms does not narrow the mission of education. It makes that mission defensible.