Preface: What This Document Is

The educational philosophy document that circulates under my name is a distillation of this larger work. It runs a little over three thousand words because search committees rarely have the time for deep exploration, and anything that calls itself a philosophy that cannot be outlined at that length probably is not a philosophy but more of a manifesto.

This is the long version.

It exists because every compressed claim in the short version rests on a body of evidence, a set of texts, and a run of arguments and experiences that anyone who cares to look is entitled to inspect, and because I promised myself when I wrote the short version that I would not assert anything there I could not defend here. This is just as much an open research diary as it is an academic document, but it is one that I share openly, in the hopes that regular revisitation will lead to refinement.

My philosophy stands on two historical methods and departs from both at a single point. The methods are the modus parisiensis, the organizational style the University of Paris developed across the fifteenth and sixteenth centuries, and the Ratio Studiorum of 1599, the Jesuit plan of studies that codified it. The point of departure is the end each was built to serve. I would keep the methodology and most of the practice, adopt a less constrained Aristotle than either permitted, and put rigorous ethical inquiry where the formational and theological end used to sit. Everything else in this document is an account of that architecture, a defense of the substitution, or an audit of the evidence for the practices that follow from it.

1The Method Worth Inheriting

The modus parisiensis was never a formally published methodology. It is a descriptive label, applied retrospectively by historians, for the way the colleges of the University of Paris organized teaching in the century before the Jesuits existed, and the best short definition I know is Grendler’s: the phrase “meant teaching practices: graduated progress in learning texts characterized by numerous drills, exercises, and disputations patterned on classroom practices used in the Paris colleges at which Ignatius Loyola and almost all the other original Jesuits studied” (Grendler, 2014). Codina’s 1968 monograph remains the standard scholarly account of the method’s origins, and his chapter in the Fordham anniversary volume is the most accessible English summary of what the Jesuits actually took from Paris (Codina, 2000). O’Malley, whose student I had the privilege of being, wrote the history of the Society’s first decades, and it carries a section titled, pointedly, “Transcending the Modus Parisiensis,” a reminder that the Jesuits did more than borrow the method; they systematized it past anything Paris had done (O’Malley, 1993).

The practices the label covers are concrete. Students were divided into graded classes by attainment rather than age, and a class was understood as a body of work to be mastered rather than a span of time to be served, so a capable student moved faster and a struggling one was not promoted on the calendar. Promotion ran by examination. Large classes were subdivided into groups of ten, the decuriae, each under a stronger peer, the decurio, who heard the recitations of the ten and reported to the teacher. Schwickerath, writing in 1903 with the Ratio open in front of him, traces the decuria back through the Brethren of the Common Life and confirms that “at the head of each decuria was a decurio, to whom his ten subjects had to recite their lessons,” and quotes Polanco, the Society’s secretary and himself a Paris man, writing of the early colleges at Messina and Vienna that exercises “were added to the lectures after the model of those of Paris (more parisiensi)” (Schwickerath, 1903). Above all there were the exercitia, the drills, repetitions, compositions, and disputations in which a student demonstrated mastery aloud and in public. The Jesuits encountered this system as students, adopted it as teachers, refined it through the draft Ratio of 1586 and the revision of 1591, and in 1599 published it as the Ratio atque Institutio Studiorum Societatis Iesu, at which point, as Grendler puts it, the Society “had a fixed curriculum and detailed pedagogical methods to follow” (Grendler, 2014). The 1599 text survives in early printings, and the 1606 Roman edition is digitized and readable (Societas Iesu, 1606). Farrell’s 1970 English translation is the working text for anyone who reads the Ratio in this country (Farrell, 1970), and Pavur’s 2005 translation is the more recent and more carefully annotated one (Pavur, 2005).

The four instruments

Four of the Ratio’s instruments still hold up, in my estimation, inside a modern pedagogical framework.

The praelectio, the prelection, has a narrow and superb purpose. The teacher reads the whole passage aloud without interruption and gives its sense briefly in the vernacular, then works it: connects it to the previous lesson, opens its vocabulary and structure, names its difficulties, and stops. The prelection exists so that the student can then work the text alone, and everything about its design, from its brevity to its refusal to do the student’s work on the student’s behalf, serves that release. In the early years it addresses syntax; in the later years, the humanities and rhetoric classes, it moves to argument, allusion, and the fitness of an author’s style. Farrell, in his introduction, quotes Donohue on why this discipline decays in practice: “It is only too easy, unfortunately, to neglect this work of preparing students for individual study since it is always easier to tell them what than to teach them how” (Donohue, 1963, pp. 150-151, as cited in Farrell, 1970, p. vii).

The repetitio was layered. There was a repetition immediately following the prelection, a repetition the next day, a weekly review, and a monthly review, a pattern the Ratio applied in every grade; Schwickerath’s summary for the Catholic Encyclopedia says simply that “frequent repetitions (daily, weekly, and monthly) were carried on in all grades” (Schwickerath, 1911). The memory work was real, and it was the smaller part. Repetition under a teacher’s eye is a diagnostic instrument — the place where a student’s weakness becomes visible before an examination makes it expensive, and where the act of restating a thing generates the next question about it. Anyone who has taught a student with a language-based learning disability knows that the number of repetitions a given student needs says nothing about that student’s intelligence. It is a fact about the student, like a shoe size, and the Ratio’s structure assumed the fact and built for it four centuries before the field had a name for the students who need six repetitions where their classmates need two.

The disputatio gave argument formal roles. There was a defendant who held a set of theses, objectors who attacked them, and a presiding professor who, in the Ratio’s instruction, “should appear to be arguing with both defendant and objector” and who corrected or elaborated the arguments as they went (Farrell, 1970, Common Rules of Professors of the Higher Faculties, no. 18). The number of theses was capped with bureaucratic precision by the Prefect of Studies: no more than fifty for a general defense, no more than twenty for a particular one, twelve or fifteen for the monthly disputations, eight or nine for the weekly (Farrell, 1970, Rules of the Prefect of Studies, no. 14). Class disputations were to run two hours. The Ratio regarded the length of a disputation as a thing to be legislated, because an argument cut short has only been interrupted.

The concertatio, the honesta aemulatio, put students in contest with one another, individually or in teams, and its most striking provision is the one about rank. The lower classes were organized on a mock-military footing with officers and privates, and the rule reads that “a private may match his skill with an officer, and if he comes off the victor, he should be given the rank of the defeated officer or be awarded another prize or symbol of victory as the dignity of the class and local circumstances dictate” (Farrell, 1970, Common Rules for Teachers of the Lower Classes). Contests, Schwickerath confirms, “were frequent” (Schwickerath, 1903). Rank was earned in public and lost in public.

Taken together, these four are an operating system for a school that expects every student to perform understanding aloud, repeatedly, under standards, and to be promoted when the performance is adequate and not when the calendar says it is time to move students on to the next grade level at the end of a school term.

What the Ratio is not

Several terms now treated as the Jesuit educational inheritance are not in the 1599 text. Cura personalis is the most common. Geger, who has done the philological work, finds that its “earliest usage appears to come from Fr. Wladimir Ledóchowski,” Superior General from 1915 to 1942, in a 1934 Instructio sent to the American Jesuit colleges, and he is explicit that the phrase is “an expression of recent vintage” (Geger, 2014). Magis, used as a noun with the definite article, is later still: Geger traces “the magis” to writings of Karl Rahner in the early 1960s and concludes that “Ignatius and the early Jesuits never used it to denote an element of their spirituality” (Geger, 2012). The five-part Ignatian Pedagogical Paradigm of context, experience, reflection, action, and evaluation dates to a 1993 document from the International Commission on the Apostolate of Jesuit Education, which itself describes the Ratio as the earlier code and presents the paradigm as a modern practical approach (ICAJE, 1993). All three are reasonable ideas. None of them is the method I am adapting, and I do not claim them as such.

2Regearing the Ratio

The Ratio subordinated the entire arts course to theology as its end. In philosophy Aristotle was the prescribed author, and the professor was instructed to follow him, with the standing exception that, as Schwickerath summarizes the rule, “those of his opinions which were contradictory to revealed truths were refuted,” and with the further instruction “to speak respectfully of St. Thomas Aquinas and to follow him whenever possible” (Schwickerath, 1903). The Ratio’s Aristotle was therefore an Aristotle under supervision of a church redactionist: read closely, disputed vigorously, and overruled at the boundary of Christian doctrine. Remove the supervision and you have a different object from anything Paris or the Jesuits arrived at. I retain the method and the curricular architecture. I change the end. Nobody should mistake that for a restoration.

The historical warrant for the substitution comes from the Ratio’s own disciplinary boundaries. The plan placed moral philosophy in its own course within the philosophy sequence, taught from the Nicomachean Ethics as a philosophical text with its own repetitions and its own disputation theses, and the rules for that chair kept the professor to the principal heads of moral science and away from theological questions, which belonged to a different faculty (Farrell, 1970, Rules for the Professor of Moral Philosophy). Schwickerath confirms that the philosophy course “comprised not only philosophy properly so called, but also mathematics and natural sciences,” with ethics in the third year beside psychology and metaphysics (Schwickerath, 1903). A rigorous ethics sequence running on Aristotelian lines as a distinct academic subject is therefore continuous with the plan’s own partitioning of knowledge, even where it is discontinuous with the plan’s purpose. The Jesuits built a room for ethics and kept it under the theology faculty’s roof. I am proposing to make that space the actual aspirational goal of the whole pedagotical framework.

Eudaimonia as a demanding aim

The end I would substitute for theology is eudaimonia, understood as Aristotle understood it. I first came to understand the origins of this practice chasing down the references in The Autobiography of Benjamin Franklin wherein he comed to the conclusion that achieving moral perfection a task “of more difficulty than I had imagined”. The thirteen personal virtues he espouses in that work contain the echoed of Arisototle, and not the theologically hemmed-in version.

In the function argument of Book I the human good “turns out to be activity of soul in accordance with virtue, and if there are more than one virtue, in accordance with the best and most complete,” and, he adds, “in a complete life,” since a day or a short time does not make a person blessed (Aristotle, trans. Ross, I.7). The virtues of character are acquired by habituation, which is the whole force of the epigraph: “we become just by doing just acts, temperate by doing temperate acts, brave by doing brave acts,” and moral virtue “comes about as a result of habit,” the word itself, ethike, formed by a slight variation from ethos (Aristotle, trans. Ross, II.1). The apparatus is steered by phronesis, practical wisdom, which Kraut describes as the capacity that makes the goal right and that involves a trained perception of particulars rather than the mechanical application of rules (Kraut, 2026). Kraut’s account of the two stages of ethical development, habituation in childhood followed by the acquisition of practical reasoning, is the structure of a school divided into lower and upper divisions, and I do not think that is a coincidence.

This is a hard aim. It makes character a matter of repeated action under judgment, which means a school cannot discharge its ethical obligation by posting values on a wall or in a mission statement and asking students to affirm them; it has to build the conditions under which just, temperate, and brave acts are required and assessed. And it holds that a person can be wrong about how to live and can be reasoned toward being less wrong, a claim most contemporary schools would be loathe to put forward as policy. Julia Annas’s account of virtue as intelligent practical skill, learned the way farming or building or music is learned, by drilled practice under a teacher whose corrections one eventually internalizes, is the modern philosophical statement of what the Ratio’s repetitions and disputations were doing to a student’s mind (Annas, 2011). The Jubilee Centre’s framework, the most developed institutional attempt to put Aristotelian character education into schools, defines character as “a set of personal traits or dispositions that produce specific moral emotions, inform motivation, and guide conduct,” distinguishes moral, intellectual, civic, and performance virtues, and makes practical wisdom the integrative meta-virtue that adjudicates “when the demands of two or more virtues collide” (Jubilee Centre for Character and Virtues, 2022). Kristjánsson’s book-length defense of the position argues for “a coherent, plausible and up-to-date concept, retaining the overall structure of Aristotelian character education,” against both the instrumental and the amoral versions that circulate under the same name (Kristjánsson, 2015). Nussbaum, from a different direction, supplies the civic stakes: a system that treats education “as though its primary goal were to teach students to be economically productive rather than to think critically” is a system that is quietly abandoning the humanities on which self-government depends (Nussbaum, 2010).

Disputing character education

The strongest argument against putting ethical formation at the center of a school is that character education, as it has actually been practiced in American schools, is a program of behavioral compliance poorly draped in the vocabulary of virtue. Kohn’s version of this argument is the classic one: in most programs “the point is to drill students in specific behaviors rather than to engage them in deep, critical reflection,” and almost all brands of character education proceed by trying “to ‘fix the kids’” rather than to examine the institutions the kids inhabit (Kohn, 1997). Suissa’s version is sharper and more political: character education, she argues, displaces political education by individualising and psychologising problems that are structural, so that a movement which begins by wanting students to be good ends, by default, in wanting them to be quiet (Suissa, 2015).

I accept both criticisms as descriptions of what has often been done, and I think they are the best reason to insist on the Aristotelian version over the pamphlet version. Kohn’s objection lands on programs that specify behaviors and reward them. It misses a program whose central activity is disputation, in which a student is required to hold a thesis about how to live against objectors who are trying to break it, in front of a presider whose job is to argue with both sides. Suissa’s objection lands on programs that treat virtue as a private possession. It misses a program in which civic virtue is one of four kinds, in which the question of what is worth building is asked in public, and in which religion, politics, and history are the material ethical inquiry works on. Habituation and compliance are the difference between a school that produces judgment and a school that produces obedience, and the way to keep a school on the right side of that line is to make the judgment visible: in argument, in performance, and in the assessed record of what a student did when a hard case arrived.

If closer scrutiny of an ethics program reveals a compliance program, Kohn and Suissa are right about it and it should be shut down. If closer scrutiny reveals students reasoning aloud about cases they did not anticipate, the program is doing what the Ethics was written to do.

3The Refusal of Unexamined Dogma

I hold that religious practice belongs to the family and the individual, and that a school’s proper work is study. This is a claim about institutional role, and it carries no judgment about belief or the lack of it. Where the Ratio placed spiritual formation at the center and made the arts course its servant, I would place ethical inquiry at the center and treat religion as one of the objects that inquiry examines. The two positions share a premise, which is that religion matters enough to organize a curriculum around. They differ on what the school is for.

The constitutional footing for a school that studies religion without practicing it is older than most of the arguments against it. The Supreme Court, in the same decision that ended devotional Bible reading in public schools, wrote that “nothing we have said here indicates that such study of the Bible or of religion, when presented objectively as part of a secular program of education, may not be effected consistently with the First Amendment,” and that the Bible “is worthy of study for its literary and historic qualities” (Abington School District v. Schempp, 1963, at 225). Independent schools are not bound by that holding, which is why they should read it: it describes what a school that takes religion seriously can do without becoming an arm of a house of worship.

The developmental argument

The academic study of religion is therefore required in my curriculum, and required most of all for the students to whom faith matters most. The argument is developmental. A person’s outward expression of self changes as they acquire autonomy with age, and we call that maturation. A palate grows more discriminating as it encounters more of the world, and we call that refinement. For a student whose spirituality is a deep component of identity, that component deserves the same trajectory: exposure to the breadth of world faiths, to their histories, to their internal disagreements, and to what is established about the world, so that the belief can develop sophistication, withstand scrutiny, survive the expansion of factual knowledge without fracturing, and come into alignment with the humane ethical commitments every major tradition claims for itself. A faith that has never met a hard question is not protected by that innocence. It is exposed by it, and the exposure usually arrives in the years after a school could have done anything about it.

The American Academy of Religion states the classroom posture as well as anyone has, and I quote it in full because the parallel construction is doing my work for me: “The school strives for student awareness of religions, but does not press for student acceptance of any religion. The school sponsors study about religion, not the practice of religion. The school may expose students to a diversity of religious views, but may not impose any particular view. The school educates about all religions, it does not promote or denigrate religion” (American Academy of Religion, 2010, p. 8). Its three premises for the discipline are equally useful to a faculty: religions are internally diverse, religions are dynamic, and religions are embedded in culture (American Academy of Religion, 2010, p. 12). A student who understands those three things is protected against the flattest and most dangerous errors about other people: that a tradition speaks with one voice, that it has always said what it says now, and that it can be understood apart from the histories and economies and languages it lives inside. Moore, who chaired the task force that wrote those guidelines, built the underlying case in her earlier book, where religious illiteracy is diagnosed as a condition that fuels culture wars, curtails historical understanding, and promotes religious and racial bigotry, and where the remedy is a cultural studies approach that treats religion as a phenomenon to be studied academically rather than devotionally (Moore, 2007).

The objection that neutrality is a fiction

The most serious argument against my position comes from Warren Nord and his adherents, and it is a better argument than anything the prayer-in-schools lobby has produced. Nord argued at book length that American public education is not neutral about religion but hostile to it by omission, that “educators in public institutions exclude religious points of view” from the curriculum so completely that a student can pass through twelve years of schooling without encountering a religious account of anything, and that a truly liberal education has “powerful philosophical, political, moral, and constitutional reasons” to take religion seriously across the curriculum rather than quarantining it in an elective (Nord, 1995). Nord and Haynes made the practical version of the case for ASCD, rejecting the idea of schools as “religion-free zones” and arguing that religious perspectives belong inside the teaching of history, literature, science, and economics wherever those perspectives are part of the real conversation (Nord & Haynes, 1998).

I think Nord is mostly right, and I think his argument supports my position. The claim that omission is not neutrality is correct. A school that never lets a student hear a serious religious argument about the good life, or about the origin of moral obligation, or about what is owed to the stranger, has taken a side while pretending to referee. But the remedy Nord proposes, religion taken seriously across the curriculum as a set of perspectives with standing, is study. It is what the AAR guidelines describe and what I am requiring. Where I part from Nord is the step some of his readers take on his behalf, that a school which studies religion seriously must also form students in one. A curriculum in which religious arguments are heard and weighed, and a curriculum in which one of them is presupposed, differ in the way a school differs from a seminary, and that difference survives Nord’s critique intact. The critique tells me that a school which excludes religion has taken a side. It has nothing to say against a school which examines it.

The governing principle

The principle that governs this framework reaches well past religion, and I state it in its general form so that no reader mistakes it for a position on religion alone: Never adhere to unexamined dogma. It applies to religion in the same way, and for the same reasons, that it applies to unexamined assumptions about how physics works or how a sentence is diagrammed. Belief received and never inspected is fragile the way an unexamined scientific misconception is fragile: it fails on contact with a case it did not anticipate, and it fails hardest when the stakes are highest. Kant framed enlightenment as a person’s emergence from a minority that is self-incurred, a condition maintained not by the absence of understanding but by the absence of resolve to use it, and gave the movement its motto, “Dare to know! Have the courage to use your own understanding” (Kant, 1784/1996; open text at Kant, 1784, trans. Smith). Contemporary philosophy of education grounds the same duty in respect for students as rational beings. The current Stanford Encyclopedia treatment, rewritten by Curren in 2025, presents Siegel’s ideal that education should enable individuals to be “appropriately moved by reasons,” treats autonomy as the “ability to make and act on well-informed and well-thought-out judgments about how to live,” and lays out the live debate over whether the state may require autonomy-facilitating education against the wishes of religious communities (Curren, 2025). Callan’s answer to that debate, that a liberal democracy must educate for its own constitutive ideals while remaining open to a diversity of belief that may threaten them, is the most careful treatment of the tension I know (Callan, 1997), and Brighouse’s shorter book makes autonomy in value judgments one of a child’s core interests, an interest that schooling exists to serve and that a school can betray by deferring too completely to the home (Brighouse, 2006).

My objection is to the unexamined, wherever it turns up in a curriculum. A tradition put under serious study by a student who holds it will more often deepen than dissolve, and a school that trusts its students with that examination is paying them the compliment of taking their beliefs seriously — which is more than the school that shields them can say.

4Inquiry, and What the Evidence Actually Says

From Socrates to the trial

The Socratic posture, the disciplined use of questions to expose what a learner actually understands, is the closest thing teaching has to a universal instrument, and it is the direct ancestor of the prelection and the disputation. The research literature calls it dialogic teaching, and I use the research term with faculty because it is the term with outcome data attached. A philosophy of education that cannot survive contact with its own evidence base is a liturgy, and I would not have spent this many hours on a liturgical exegesis.

The largest trial of this practice at scale was in England. The Education Endowment Foundation tested Alexander’s dialogic teaching model in a randomized design across 76 primary schools with 4,985 Year 5 pupils at randomization, delivered over twenty weeks, and found effect sizes of +0.15 in English, +0.12 in science, and +0.09 in mathematics, which the EEF’s months-of-progress conversion renders as two additional months in English and science and one in mathematics (Jay et al., 2017). The pupils eligible for free school meals made two additional months’ progress in all three subjects.

The caveats travel with the result. The confidence interval for English touched zero, at p = .05; only the science estimate excluded it cleanly. A 2019 addendum tracked the same children to their Key Stage 2 national tests and found effect sizes of +0.03 in reading and +0.06 in mathematics, with the evaluators concluding that the analyses “provide no definitive evidence that Dialogic Teaching schools made additional progress” on those measures (Jay et al., 2019). Alexander’s response is the one an honest program designer should make: the trial “relied on standardised tests in English, mathematics and science as sole measures of the programme’s educational outcomes,” and, “lacking the essential and agreed measure of spoken language it was unable to track the intervention through to the transformation of students’ talk,” even though the video record showed that teacher and student talk had in fact changed (Alexander, 2018). I find that argument persuasive and I also notice that it is the argument every intervention makes when the outcome measure disappoints, which is why I hold it lightly.

I would caution program directors about the Philosophy for Children literature with the same firmness, and for a plainer reason. The 2015 EEF evaluation, in 48 schools with 3,159 pupils, reported effect sizes of +0.12 in reading and +0.10 in mathematics, two months in each, with larger gains for disadvantaged pupils (Gorard et al., 2015). The number went into a thousand slide decks. The re-trial, across 198 schools with the highest security rating the EEF gives, found an effect on Key Stage 2 reading of +0.01 for all pupils and +0.02 for disadvantaged pupils, zero months in mathematics, and “no evidence” of impact on social and communication skills; it also found that 35 of the 75 intervention schools were not implementing the program at the expected level after two years (Lord et al., 2021). The first result is still quoted. The second is the one that counts, and a school leader who keeps citing the first after the second has been published is making a choice about evidence, and everything that leader says afterward should be weighed with that choice in mind.

What the older literature adds, and what a meta-analysis subtracts

The trial evidence is recent. The observational and quasi-experimental literature is older and, on the whole, kinder. Applebee and colleagues, across 64 middle and high school English classrooms in four states, found that “discussion-based approaches were significantly related to spring performance” after controlling for fall performance, and that students exposed to “discussion-based approaches in the context of high academic demands internalize the knowledge and skills necessary to engage in challenging literacy tasks” (Applebee et al., 2003). The load-bearing phrase is the one about high academic demands. Discussion in a classroom without them is conversation. Michaels, O’Connor, and Resnick supplied the structure that makes discussion academically demanding, the accountable talk framework in which classroom argument is held to “accountability to the learning community,” to “accepted standards of reasoning,” and to “knowledge,” which is a modern statement of the disputation’s three constraints (Michaels et al., 2008). The AERA volume that gathers the field’s evidence under the title of socializing intelligence through talk is the best single collection (Resnick et al., 2015), and Mercer and Littleton’s sociocultural account, in which “the quality of classroom talk improves the quality of children’s thinking and educational attainment,” is the theoretical backbone (Mercer & Littleton, 2007).

Against all of that stands the meta-analysis. Murphy and colleagues, reviewing the effects of classroom discussion on text comprehension, found that discussion approaches produced “strong increases in the amount of student talk and concomitant reductions in teacher talk,” but that “few approaches to discussion were effective at increasing students’ literal or inferential comprehension and critical thinking and reasoning,” with effects moderated by study design, by the outcome measure, and by student ability (Murphy et al., 2009). I read that finding as a specification. Talk that changes the distribution of who is speaking without changing the standard to which the speaking is held produces more talk. The Ratio’s instruments measured theses defended, objections answered, and rank earned, never the quantity of student speech, and the meta-analysis is telling me, in the language of effect sizes, what Donohue told Farrell in 1963: the easy thing is to let students talk, the hard thing is to teach them how.

Engagement, measured and unmeasured

Fredricks, Blumenfeld, and Paris gave the field its working definition of engagement, behavioral, emotional, and cognitive, and in the same review described the construct as “presumed to be malleable, responsive to contextual features, and amenable to environmental change” while conceding that “the potential contribution of the concept of school engagement to research on student experience has yet to be realized” (Fredricks et al., 2004). Reschly and Christenson catalogued the “jingle, jangle, and conceptual haziness” of a field in which the same terms name different constructs and different terms name the same one (Reschly & Christenson, 2012), and Wong and Liem, a decade later, found the problems intact: overgeneralization, jingle-jangle fallacies, ambiguity about what the student is engaged with, and a tendency to define engagement in terms of the outcomes it is supposed to predict, which is circular (Wong & Liem, 2022).

With those cautions stated, the correlational picture is consistent. Across 69 independent studies and 196,473 students, engagement correlates with achievement at r = .35 for behavioral engagement, .25 for cognitive, and .22 for emotional, with an overall correlation of .27, every estimate significant and every one moderated by how engagement was reported and where the study was done (Lei et al., 2018). Those are associations. Nobody has randomized engagement.

The national picture is poor. Gallup’s 2024 survey of Gen Z, with 2,317 enrolled K-12 students in the sample, found that 18 percent strongly agree their teachers make schoolwork feel important, 17 percent strongly agree they have recently learned something interesting at school, and 14 percent strongly agree that their coursework challenges them in a good way, which Gallup summarizes as fewer than two in ten students strongly agreeing that classwork “feels important, interesting, challenging or aligned with their talents” (Hrynowski, 2024). So the defensible claim is this: engagement is among the few school-alterable conditions with a consistent relationship to outcomes; the construct is loose enough that a school must decide what it means by the word before it measures anything; and a school that cannot say how it knows whether its students are engaged is flying without instruments.

5Work That Produces Something Tangible

Modernizing the exercitium

Project-based and experiential learning are how inquiry acquires tangible returns, and they are the modern form of the exercitium. A composition in the Ratio’s rhetoric class was a made thing, judged against a standard, in front of peers who were making the same thing; nobody in that room would have mistaken it for a warm-up. Dewey made the philosophical case for continuity between experience and education, and made it against the progressive excesses of his own followers as much as against the traditionalists, which is a detail his admirers tend to forget (Dewey, 1938/1998). Kolb supplied a usable cycle, concrete experience, reflective observation, abstract conceptualization, active experimentation, which has the virtue of being teachable to a faculty in an afternoon and the vice of being taught as if the afternoon were the whole of it (Kolb, 1984). What neither of them supplied was randomized evidence, and the empirical case for project-based learning has only recently become respectable.

What the trials show

Three randomized trials carry the weight of the claims above. Krajcik and colleagues randomized 46 Michigan elementary schools and followed 2,371 third graders through a year of project-based science; the treatment students scored 0.277 standard deviations higher on a standardized science assessment, and reported higher self-reflection and collaboration besides (Krajcik et al., 2023). Saavedra and colleagues randomized 74 high schools across five large urban districts into a project-based version of two Advanced Placement courses, U.S. Government and Politics and Environmental Science, and found “positive and significant treatment effects on AP exam performance for the overall sample, within both AP courses studied, and within low- and high-income student groups” (Saavedra et al., 2022). In the first year, 35.1 percent of treatment students earned a qualifying score against 31.1 percent of controls, a four-point gain with an effect size of 0.264, rising to 7.6 points among students who actually sat the exam, with the effect among lower-income exam-takers at 0.386 (Saavedra et al., 2021). The second-year results were positive and not statistically significant, which press accounts of the study tend to omit. Duke and colleagues randomized 48 second-grade teachers in low-income districts into four project-based units and found higher growth in social studies and informational reading, no significant difference in writing or motivation, and one finding a Head of School should read twice: “greater consistency with project-based learning session plans was associated with higher growth in writing, motivation, and reading” (Duke et al., 2021). Fidelity mattered. That is an administrative finding as much as a pedagogical one. It tells a school leader where the support goes, which is to the teachers doing the implementing and not to the viewbook.

Academic inflation

A meta-analysis reports a pooled effect of roughly 0.71 standard deviations for project-based learning across two decades of studies (Chen & Yang, 2019). I treat that number with suspicion, and the reason is one a teacher can carry into a faculty meeting. Pooling small, mostly non-randomized studies, many with researcher-designed outcome measures aligned to the intervention, inflates effects, and an effect of 0.71 is more than twice the size of the effects the three randomized trials actually found. When a meta-analysis says a thing works far better than the best trials of that thing say it works, the meta-analysis is measuring the literature, not the intervention. MDRC’s review remains the fairest summary: the evidence for project-based learning is “promising but not proven,” with the positive findings concentrated in science and social studies and the weakest evidence in mathematics and literacy (Condliffe et al., 2017). That is where the three trials sit too. I will build a science and social studies program on project work with confidence and a mathematics program on it with care.

The minimal-guidance objection

The standing objection to all of this is Kirschner, Sweller, and Clark’s, and it is a good one. Reviewing the cognitive-load literature, they concluded that “minimally guided instruction is less effective and less efficient than instructional approaches that place a strong emphasis on guidance,” and they filed constructivist, discovery, problem-based, experiential, and inquiry-based teaching under the same heading of minimal guidance (Kirschner et al., 2006). The reply from Hmelo-Silver, Duncan, and Chinn was that the frame of the inquiry was wrong: problem-based and inquiry learning as actually designed “employ scaffolding extensively thereby reducing the cognitive load and allowing students to learn in complex domains,” and the studies Kirschner and colleagues cited were mostly of pure discovery, which nobody in the field defends (Hmelo-Silver et al., 2007). Both sides are right about the thing in front of them. Unguided discovery often fails, and it fails worst for the students with the least prior knowledge, a population that grows as literacy declines nationally and background knowledge atrophies with it. Heavily scaffolded project work, in which the prelection has happened, the knowledge is in place, and the project is the exercitium that demonstrates mastery instead of the vehicle that is supposed to produce it, is the thing the three trials tested, and it works. The Ratio would have found the whole controversy puzzling, because it never occurred to anyone in 1599 that a student should compose an oration before being taught to read one.

6Thinking is Connecting Knowledge to Potential

Inquiry and projects fail without knowledge. Critical thinking is domain knowledge in action, and Willingham has spent two decades arguing that it cannot be trained as a portable skill in the abstract: “the processes of thinking are intertwined with the content of thought (that is, domain knowledge),” and “critical thinking is not a set of skills that can be deployed at any time, in any context” (Willingham, 2007). The cleanest demonstration is nearly forty years old. Recht and Leslie gave 64 seventh and eighth graders, sorted into good and poor readers, a passage about a half-inning of baseball and found that “greater knowledge of the text’s subject resulted in improved recall and retention” and that there were “no interactions … between prior knowledge and reading ability”: the poor readers who knew baseball outperformed the good readers who did not (Recht & Leslie, 1988). Comprehension past the decoding stage is largely a function of what a reader already knows. Hirsch built a career on that sentence and a curriculum on the conviction that “a carefully planned, knowledge-based elementary curriculum is essential,” and his last book is the most direct statement of the case (Hirsch, 2016). Cabell and Hwang bring the argument into the primary grades with the evidence, still described by its authors as preliminary, that “knowledge building in English language arts instruction can support language and content knowledge, leading to better linguistic and reading comprehension” (Cabell & Hwang, 2020).

The science of reading and its dissenters

The consensus reading science treats systematic phonics as foundational. The National Reading Panel’s review found that systematic phonics instruction produces significant gains across student populations and grades (National Institute of Child Health and Human Development, 2000), and Castles, Rastle, and Nation’s attempt to end the reading wars puts the point at its simplest: “teaching phonics provides children with the principal means of getting from the printed form of a word to its spoken form,” after which “reading comprehension is multifaceted and requires considering the range of linguistic and cognitive processes implicated in text comprehension,” meaning background knowledge, vocabulary, and executive function (Castles et al., 2018).

Bowers reviewed twelve meta-analyses and concluded that “systematic phonics provides a small short-term benefit to spelling, reading text, and comprehension, with no evidence that these effects persist following a delay of 4-12 months,” and that the comparison conditions in the trials were rarely the whole-language or balanced approaches phonics is said to have beaten (Bowers, 2020). Wyse and Bradbury, from a meta-synthesis of 55 trials and a survey of 2,205 teachers in England, argue that “the approach to phonics and reading teaching in England is not sufficiently underpinned by research evidence” and that contextualized teaching of reading, phonics included, outperforms synthetic phonics delivered in isolation (Wyse & Bradbury, 2022).

I read both as arguments about dosage and context. Neither denies that a child must learn to map print to sound. Both deny that mapping print to sound, taught alone and taught long, is a reading program, and on that point they agree with Castles, Rastle, and Nation, whose title was chosen because the evidence supports phonics as a foundation and not as a building. For the students I have spent the last fifteen years serving, the ones for whom decoding does not arrive on its own, the foundation is non-negotiable, and it is the first course of stone a remediation program lays.

Deep literacy as the core

Deep literacy is the capacity to read something difficult and unfamiliar and come out the other side holding an understanding of it, along with new knowledge adjacent to what was read. That is the first item in the core. The others: the fundamentals of mathematics and the applied sciences. A shared body of literature that gives a community common reference, since a school whose students have read nothing in common has no common language for exploration, argument, or exposition. Formal instruction in how to conduct research, how to find, weigh, and cite what others have established, a skill the Ratio assumed in its disputants and the modern school delegates to a mish-mash of humanities-adjacent practices. Beside those should sit the applied disciplines of running a project, writing code, and building the better version of a thing, alone or in a group, which are the same competence expressed in different materials, since each is the discipline of holding a specification in mind while the material resists it.

This conviction comes out of a decade and a half inside schools built for students with language-based learning disabilities. For those students the distance between surface literacy and deep literacy is widest, and the temptation to settle for literate enough is strongest, because literate enough is measurable, reportable, and kind in the short run. Emotional protection lowers the bar; the lowered bar becomes the new standard; tomorrow’s lowered bar falls below today’s, and at every step the move is described as compassion. The Ratio’s answer to that ratchet was promotion by examination, which has its cruelties and which I would not restore whole, but which at least refused to lie to a student about where they stood. My answer is the same goal with better instruments: lower the floor as far as the student needs, and never lower the ceiling.

7Heutagogy: The Aim Past Graduation

Here the lineage pedagogy and the modern literature converge. Pedagogy has the teacher determining what is learned and how. Andragogy, in Knowles’s formulation, gives the adult learner control of the process inside a curriculum someone else frames, on the theory that adults arrive with experience to draw on and problems to solve rather than subjects to complete (Knowles, 1980). Heutagogy, as Hase and Kenyon proposed it in 2000, moves the determination itself to the learner: “it is the learner [who] determines what and how learning should take place,” and the goal shifts from competency, which “consist[s] of knowledge and skills,” to capability, which is “a holistic attribute” whose possessors “know how to learn; are creative; have a high degree of self-efficacy; can apply competencies in novel as well as familiar situations” (Hase & Kenyon, 2000). Blaschke’s review a dozen years later added the element the compressed version of this document attributed to the original, the negotiated assessment, in which “learners are highly autonomous and self-determined” and the assessment of learning is itself a matter of agreement between learner and teacher (Blaschke, 2012); Blaschke and Hase’s later chapter frames the whole as “a holistic, learner-centered approach to learning and teaching, in formal and informal situations” (Blaschke & Hase, 2016).

The prelection already pointed this way. Its entire purpose was to make a student able to work the text alone, which is a four-hundred-year-old statement of the same aim, and Donohue’s warning that it is easier to tell students what than to teach them how is the warning heutagogy issues against the teacher who cannot let go.

Heutagogy in the wild

I present heutagogy as a philosophical commitment rather than an evidence-based intervention, because I went looking for the evidence and it is thin. The most recent review of the subject is narrative, not systematic; it screened 268 records down to 22 articles, drawn from PubMed, Medline, and related databases, and found that “many studies rely on anecdotal or theoretical claims rather than robust outcome data,” with “limited empirical validation across diverse educational contexts” (Panta, 2025). The included studies come from health professions education, digital and online learning, problem-based learning, and teacher education. None of them is a school-age study. That absence is partly an artifact of where the review searched, and a second 2025 review in the career and technical education literature does discuss heutagogical framing for K-12 digital tools, though without outcome data of its own (Newfield, 2025). The K-12 empirical base is effectively absent, and I will not manufacture one.

There is also a principled objection, and it is the same one that shadowed project-based learning. Kirschner and van Merriënboer list the learner who knows best, who can be trusted to direct their own learning, among the urban legends of education, alongside the digital native and the learning style, on the grounds that novices are the people least equipped to judge what they need to learn next (Kirschner & van Merriënboer, 2013). I take the objection seriously, and it is why heutagogy is an aim and never a starting condition. A second grader is a novice and is treated as one. The argument for self-determined learning stands on reasoning about the end state: students now in our buildings will spend their working lives acquiring skills that do not yet exist in fields not yet named, and a graduate who can only perform what was taught has been prepared for a world already gone. In practice this means a gradual release of determination across the divisions, so that a fourth grader chooses within a structure, a ninth grader negotiates the structure, and a senior designs and defends an inquiry no adult needed other than to provide a response to the work completed.

8Civics, Lifelong Learning, and Perseverance

Fifty years of a stated aim

Lifelong learning has been the stated aim of international educational policy for over fifty years. The Faure commission’s 1972 report set it out as the organizing principle of a “learning society” and named its ends as scientific humanism, social commitment, creativity, and the complete person (Faure et al., 1972). The Delors commission restated it in 1996 as four pillars, learning to know, learning to do, learning to live together, and learning to be, and called the whole “a necessary Utopia in which no one’s hidden talents are left untapped” (Delors et al., 1996). The 2021 futures commission reframed it as a new social contract for education “grounded in human rights and based on principles of non-discrimination, social justice, respect for life, human dignity and cultural diversity” (International Commission on the Futures of Education, 2021). Those are normative commitments, and I cite them as such. They tell me what the international consensus says a school is for. They do not tell me whether any school has achieved it, and the utopia Delors called necessary was, by his own adjective, a utopia.

Civics 50 years later

The civic half of that inheritance is in poor condition, and the accurate version is more sobering than the alarmed one. On the 2022 NAEP civics assessment, 22 percent of eighth graders performed at or above Proficient, 69 percent at or above Basic, which leaves 31 percent below it, and the average score was two points lower than in 2018 and not significantly different from the first civics assessment in 1998 (National Center for Education Statistics, 2023a). U.S. history is worse: 13 percent at or above Proficient, 60 percent at or above Basic, so 40 percent below, an average score five points lower than 2018, nine points lower than 2014, and not significantly different from 1994 (National Center for Education Statistics, 2023b). The recent erosion concentrates among the lowest-performing students. In civics the declines since 2018 appear at the 10th, 25th, and 50th percentiles while the 75th and 90th held; in history the declines run from the 10th through the 75th percentile while only the 90th held. That distribution makes it an equity problem before it is a curricular one. The students losing civic knowledge are the students with the least of it to lose.

The remedy is known. Campbell’s review of what social scientists have learned about civic education concludes that “well-designed civic education, both formal and informal, has meaningful, long-lasting effects on the civic engagement of young people,” and, more to the point for a school serving students whose homes cannot supply the deficit, that “effective civic education can compensate for a dearth of civic resources in the home and community” (Campbell, 2019). The four dimensions he identifies, classroom instruction, extracurricular activity, service learning, and institutional ethos, are the Ratio’s dimensions with the names changed. The Educating for American Democracy roadmap, produced with federal support and organized around inquiry rather than coverage, is the most credible curricular framework now available, and its own accounting of the problem is blunt: the country spends approximately fifty dollars per student per year on STEM and approximately fifty cents on civics (Educating for American Democracy, 2021). Civics, ethics, and philosophy are the disciplines that convert capability into judgment. A student who can research anything and build anything, and who has never been required to ask what is worth building, is a hazard, whatever the transcript says.

On grit, the word I will not use

I want graduates durable enough to finish hard things, and I am wary of the vocabulary that usually accompanies that wish. The grit literature began well. Duckworth and colleagues defined grit as perseverance and passion for long-term goals, showed across six samples that it predicted outcomes over and beyond IQ and conscientiousness, and reported, in the same paper, that it accounted for an average of four percent of the variance in success outcomes and correlated at .77 with conscientiousness in one sample, two figures that should have tempered the reception and did not (Duckworth et al., 2007). The short scale that followed replicated the conscientiousness correlation at .77 and found that the Spelling Bee result was “mediated by lifetime spelling practice,” which is to say that grit predicted spelling because gritty children spelled more, not because grit is a substance (Duckworth & Quinn, 2009).

Then came the meta-analysis. Credé, Tynan, and Harms, across 584 effect sizes from 88 independent samples and 66,807 people, found that the two facets, consistency of interest and perseverance of effort, do not cohere as a single higher-order construct, that grit is very strongly correlated with conscientiousness, that its relationships with performance and retention are moderate, that “the primary utility of the grit construct may lie in the perseverance facet,” and that “interventions designed to enhance grit may only have weak effects on performance and success” (Credé et al., 2017). Credé’s follow-up for educators is more direct still: “there appears to be no reason to accept the combination of perseverance and passion for long-term goals into a single grit construct,” and there is no support for the claim that grit “is likely to be responsive to interventions” (Credé, 2018). The defense, from Jachimowicz and colleagues, argues from 127 studies and 45,485 participants that the meta-analysis mismeasured passion, and that perseverance combined with a properly measured passion attainment predicts performance better than perseverance alone (Jachimowicz et al., 2018). That is a real disagreement between serious people and I do not need to resolve it, because both sides agree on the part I need. Perseverance of effort does the predictive work.

So I will not teach grit. I will run a program demanding enough that persistence is required, coherent enough that it is rewarded, and supported enough that it is possible. That is Aristotle again. We become persevering by persevering, under conditions that make persevering both necessary and survivable, and a school that wants durable graduates builds those conditions instead of a curriculum unit about them.

9Technology as an Instrument of This Philosophy

I have spent thirty years as a technologist inside schools and mission-driven institutions, and the discipline I apply to technology is the one this document applies to everything else: an instrument is worth what it makes possible, and nothing on its own.

The titles I have held invite the wrong assumption about where my attention goes, so I will state the comparison plainly. Many colleagues I respect, here and abroad, are better engineers than I am, better programmers, better network and security practitioners. That is a wide field of narrow columns. The narrow field is the people who hold one wide column: technical depth beside educational rigor, mission-aligned practice, and a working commitment to equity in literacy. That intersection is where I have spent my career, and it is why I describe the work as technology in education and not as technology that happens to occur in a school.

Ed-Tech Skeptical

The skeptical literature on educational technology is large, and I begin with it because anyone who proposes or builds digital systems for schools should be made to. The OECD’s 2015 analysis of PISA data found no appreciable improvement in reading, mathematics, or science in the countries that had invested most heavily in information technology for education, and concluded that ensuring every child reaches a baseline of proficiency in reading and mathematics would do more for equal opportunity in a digital world than expanding access to devices (OECD, 2015). The UNESCO monitoring report of 2023, whose subtitle asks on whose terms technology enters education, found that fewer than a quarter of countries regulate smartphones in schools, that 89 percent of the 163 education technology products it reviewed could surveil children, and that the share of education interventions in the United States with strong evidence of effectiveness was in the low single digits, and it stated flatly that technology should never replace in-person, teacher-led instruction (UNESCO, 2023).

Cheung and Slavin’s meta-analysis of 74 studies and 56,886 K-12 students found that educational technology applications for mathematics “generally produced a positive, though modest, effect (ES = +0.15) in comparison to traditional methods,” with supplemental computer-assisted instruction at +0.18 and the comprehensive models that promised the most delivering the least, at +0.07 (Cheung & Slavin, 2013). Reich, after a decade of studying learning at scale, put the finding in one sentence: “Technology is a phenomenal tool in the right hands, but no killer app will shortcut the hard road of institutional change” (Reich, 2020).

Every education technology pitch of the last decade has invoked Bloom, whose 1984 paper reported that students tutored one-to-one performed about two standard deviations above students in conventional classrooms and set the field the problem of finding group methods as effective as a tutor (Bloom, 1984). The number was always an outlier. The most rigorous recent synthesis of experimental tutoring studies finds “consistent and substantial positive impacts on learning outcomes, with an overall pooled effect size estimate of 0.37 SD,” which is large by the standards of anything else a school can do and less than a fifth of the figure the pitch decks quote (Nickow et al., 2020). A school leader who knows the difference between two sigma and 0.37 will save their institution a great deal of money, or at least the embarrassment of having promised what the money could not buy.

The specific good

Against that record, the specific good technology delivers is access to the demanding core described above, and for the population I serve the evidence is unusually clean. Wood and colleagues’ meta-analysis of text-to-speech and read-aloud tools for students with reading disabilities, across 22 studies, found a mean effect on comprehension of d = .35 with a confidence interval from .14 to .56; the tools “may assist students with reading comprehension,” and a student whose decoding fluency lags can therefore meet a text of full complexity with the right support under it (Wood et al., 2018). MacArthur and Cavalier’s study of dictation and speech recognition is more telling still, because of who it did not help: “for students with LD, essays dictated using speech recognition were better than handwritten essays, and essays dictated to a scribe were even better,” while students without learning disabilities showed no difference across modes (MacArthur & Cavalier, 2004). That asymmetry is the signature of an accommodation. A shortcut helps everyone. The tool removed a barrier for the students who had one and did nothing for the students who did not, and that is the profile a school should demand before it adopts anything. The Universal Design for Learning guidelines, in their current revision, state the goal as ensuring that “all learners are able to access and participate in meaningful, challenging learning opportunities,” and the adjective doing the work in that sentence is challenging (CAST, 2024).

So a student whose decoding lags can still meet a text of full complexity. A student who cannot yet produce written output at grade level can still construct an argument and be held to its quality. A student who requires six repetitions can receive them without public cost, which is the Ratio’s decuria with the shame removed. Each of these removes a barrier between a learner and rigorous content, which is the practical form of a commitment to flourishing for students whose paths to it are obstructed. The failure mode is the mirror image: technology used to lower the demand instead of removing the barrier, producing students who complete work they did not do. In practice the difference is visible from the classroom doorway, and holding a faculty to it is leadership work before it is procurement work.

LLM solutioning

The same test governs LLM deployments, popularly called AI, in education. Having designed and deployed production systems in a school and written the policy and training that governed their use, I am more cautious now than I was at the start of this boom, and the trial evidence of the last two years has given the caution its currnt edge.

Bastani and colleagues ran a field experiment with nearly a thousand high school mathematics students and gave some of them an unconstrained model and some a model built with tutoring guardrails that withheld answers. Practice performance rose 48 percent with the unconstrained model and 127 percent with the guardrailed one. When access was removed for the exam, the students who had used the unconstrained model scored 17 percent lower than students who had never had it, while the guardrailed group’s harm was largely eliminated (Bastani et al., 2025). That experiment is my policy. A tool that returns a correct answer while leaving a student less able to reach it alone has not helped that student, and in schools serving learners who already struggle, that is the entire risk.

The mechanism has a name in the cognitive literature: cognitive offloading, “the use of physical action to alter the information processing requirements of a task so as to reduce cognitive demand,” which is a benefit when the demand is incidental and a theft when the demand is the lesson (Risko & Gilbert, 2016). A widely circulated preprint on essay writing with a language model reports weaker neural connectivity, lower self-reported ownership, and an inability to quote one’s own work among the model-assisted group; it is a preprint with 54 participants and I cite it as suggestive and nothing more (Kosmyna et al., 2025).

On the other side of the ledger, Kestin and colleagues, in a randomized trial with 194 Harvard physics students, found that a tightly constrained AI tutor produced learning gains of about 0.63 standard deviations over an in-class active learning session, in less time and with higher reported engagement (Kestin et al., 2025). The World Bank’s Nigerian pilot, six weeks of after-school sessions with a general-purpose assistant, found gains of 0.31 standard deviations on a composite assessment and 0.23 on English, which the authors describe as the equivalent of between one and a half and two years of ordinary schooling, with larger effects for girls and for higher-performing students (De Simone et al., 2025). Those are large effects, and the second one complicates my framing, because the students who gained most were not the students with the greatest obstacles.

I read both studies as confirmations of the guardrail finding. The Harvard tutor was designed by physicists to withhold answers and make students work; the Nigerian sessions were teacher-supervised and structured. Neither is a student alone with an answer engine at eleven at night, and that is the condition a school policy has to govern.

The test I apply to any technology is whether it increases what a student can eventually do without it. Applied consistently, that test resolves most of the debates a school will have about devices, about generative tools, and about accommodation, because it converts a question of policy into a question of purpose. Text-to-speech for a student with dyslexia passes, because the student reads harder texts and, in the fullness of time, reads more. An answer engine for a student with a homework problem fails, because the student completes the problem and learns to complete it that way. A guardrailed tutor passes conditionally, and the condition is the guardrail. There is nothing in that test that any member the faculty cannot apply on a Tuesday afternoon, and nothing in it that requires a technologist, which is the point of writing it down.

Final Analysis

A graduate of a school run on these principles should leave with a nimble mind, geared toward acquiring whatever knowledge and skill a goal turns out to require, furnished with enough of the humanities to make the goals worth pursuing, and practiced enough in difficulty to see them through. The method supplies the discipline. The core supplies the knowledge. Ethical inquiry supplies the judgment. Heutagogy is the point at which the school lets go.

Read back through the sections, the architecture is one structure seen from several sides. The prelection and the release of determination are the same act at two ages. The disputation and the senior inquiry are the same assessment at two levels of stake. The decuria and the accommodation are the same recognition, that students need different numbers of repetitions, expressed in 1599 as a rank and in 2026 as a tool. The concertatio’s earned rank and the negotiated assessment are the same bargain, struck with the student instead of over them. The Ratio’s refusal to promote by calendar and my refusal to call a student literate enough are the same refusal. What changed in four hundred years is the end the apparatus serves, and how far it is allowed to carry a student.

I hold these positions with the confidence of someone who has implemented programs, and the humility of someone who has watched some of those implementations fall short. The evidence behind them is uneven, and this document has tried to illustrate how uneven and where: strong for phonics and knowledge, moderate and correlational for engagement, promising and subject-dependent for project-based learning, disputed but convergent on perseverance, clean for assistive technology and newly clarifying for LLM solutioning, and, for heutagogy, absent beyond the philosophy.

The March Hare told Alice she should say what she meant, and Alice answered that she meant what she said, which was the same thing. My educational philosophy statement says what I mean. This one is the document where I find out whether I mean it. I leave all this out in the open, with the sources on the table for anyone to tell me where I have overlooked, overstated, or plainly misunderstood. The Hatter had the last word in that exchange, and he was right.

Not the same thing a bit.

Master Reference List

1. Abington School District v. Schempp, 374 U.S. 203 (1963). https://supreme.justia.com/cases/federal/us/374/203/
2. Alexander, R. (2018). Developing dialogic teaching: Genesis, process, trial. Research Papers in Education, 33(5), 561-598. https://doi.org/10.1080/02671522.2018.1481140
3. American Academy of Religion. (2010). Guidelines for teaching about religion in K-12 public schools in the United States. https://aarweb.org/common/Uploaded%20files/Publications%20and%20News/Guides%20and%20Best%20Practices/AARK-12CurriculumGuidelinesPDF.pdf
4. Annas, J. (2011). Intelligent virtue. Oxford University Press. https://global.oup.com/academic/product/intelligent-virtue-9780199228782
5. Applebee, A. N., Langer, J. A., Nystrand, M., & Gamoran, A. (2003). Discussion-based approaches to developing understanding: Classroom instruction and student performance in middle and high school English. American Educational Research Journal, 40(3), 685-730. https://doi.org/10.3102/00028312040003685
6. Aristotle. (n.d.). Nicomachean ethics (W. D. Ross, Trans.). The Internet Classics Archive, Massachusetts Institute of Technology. http://classics.mit.edu/Aristotle/nicomachaen.html (Original work published ca. 350 B.C.E.)
7. Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakcı, Ö., & Mariman, R. (2025). Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences, 122(26), Article e2422633122. https://doi.org/10.1073/pnas.2422633122
8. Blaschke, L. M. (2012). Heutagogy and lifelong learning: A review of heutagogical practice and self-determined learning. The International Review of Research in Open and Distributed Learning, 13(1), 56-71. https://doi.org/10.19173/irrodl.v13i1.1076
9. Blaschke, L. M., & Hase, S. (2016). Heutagogy: A holistic framework for creating twenty-first-century self-determined learners. In B. Gros, Kinshuk, & M. Maina (Eds.), The future of ubiquitous learning: Learning designs for emerging pedagogies (pp. 25-40). Springer. https://doi.org/10.1007/978-3-662-47724-3_2
10. Bloom, B. S. (1984). The 2 sigma problem: The search for methods of group instruction as effective as one-to-one tutoring. Educational Researcher, 13(6), 4-16. https://doi.org/10.3102/0013189X013006004
11. Bowers, J. S. (2020). Reconsidering the evidence that systematic phonics is more effective than alternative methods of reading instruction. Educational Psychology Review, 32(3), 681-705. https://doi.org/10.1007/s10648-019-09515-y
12. Brighouse, H. (2006). On education. Routledge. https://www.routledge.com/On-Education/Brighouse/p/book/9780415327909
13. Cabell, S. Q., & Hwang, H. (2020). Building content knowledge to boost comprehension in the primary grades. Reading Research Quarterly, 55(S1), S99-S107. https://doi.org/10.1002/rrq.338
14. Callan, E. (1997). Creating citizens: Political education and liberal democracy. Oxford University Press. https://global.oup.com/academic/product/creating-citizens-9780199270194
15. Campbell, D. E. (2019). What social scientists have learned about civic education: A review of the literature. Peabody Journal of Education, 94(1), 32-47. https://doi.org/10.1080/0161956X.2019.1553601
16. CAST. (2024). Universal Design for Learning guidelines version 3.0. https://udlguidelines.cast.org
17. Castles, A., Rastle, K., & Nation, K. (2018). Ending the reading wars: Reading acquisition from novice to expert. Psychological Science in the Public Interest, 19(1), 5-51. https://doi.org/10.1177/1529100618772271
18. Chen, C.-H., & Yang, Y.-C. (2019). Revisiting the effects of project-based learning on students’ academic achievement: A meta-analysis investigating moderators. Educational Research Review, 26, 71-81. https://doi.org/10.1016/j.edurev.2018.11.001
19. Cheung, A. C. K., & Slavin, R. E. (2013). The effectiveness of educational technology applications for enhancing mathematics achievement in K-12 classrooms: A meta-analysis. Educational Research Review, 9, 88-113. https://doi.org/10.1016/j.edurev.2013.01.001
20. Codina, G. (2000). The “modus Parisiensis.” In V. J. Duminuco (Ed.), The Jesuit Ratio Studiorum: 400th anniversary perspectives (pp. 28-49). Fordham University Press. https://ecommons.luc.edu/ignatianpedagogy/197/
21. Condliffe, B., Quint, J., Visher, M. G., Bangser, M. R., Drohojowska, S., Saco, L., & Nelson, E. (2017). Project-based learning: A literature review (Working paper). MDRC. https://www.mdrc.org/sites/default/files/Project-Based_Learning-LitRev_Final.pdf
22. Credé, M. (2018). What shall we do about grit? A critical review of what we know and what we don’t know. Educational Researcher, 47(9), 606-611. https://doi.org/10.3102/0013189X18801322
23. Credé, M., Tynan, M. C., & Harms, P. D. (2017). Much ado about grit: A meta-analytic synthesis of the grit literature. Journal of Personality and Social Psychology, 113(3), 492-511. https://doi.org/10.1037/pspp0000102
24. 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/
25. De Simone, M., Tiberti, F., Barron Rodriguez, M., Manolio, F., Mosuro, W., & Dikoru, E. J. (2025). From chalkboards to chatbots: Evaluating the impact of generative AI on learning outcomes in Nigeria (Policy Research Working Paper No. 11125). World Bank. http://hdl.handle.net/10986/43212
26. Delors, J., Al Mufti, I., Amagi, I., Carneiro, R., Chung, F., Geremek, B., Gorham, W., Kornhauser, A., Manley, M., Padrón Quero, M., Savané, M.-A., Singh, K., Stavenhagen, R., Suhr, M. W., & Nanzhao, Z. (1996). Learning: The treasure within. Report to UNESCO of the International Commission on Education for the Twenty-first Century. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000109590
27. Dewey, J. (1998). Experience and education (60th anniversary ed.). Kappa Delta Pi. (Original work published 1938) https://archive.org/details/experienceeducat0000dewe
28. Donohue, J. W. (1963). Jesuit education: An essay on the foundations of its idea. Fordham University Press. https://books.google.com/books/about/Jesuit_education.html?id=v_RXAAAAMAAJ
29. Duckworth, A. L., Peterson, C., Matthews, M. D., & Kelly, D. R. (2007). Grit: Perseverance and passion for long-term goals. Journal of Personality and Social Psychology, 92(6), 1087-1101. https://doi.org/10.1037/0022-3514.92.6.1087
30. Duckworth, A. L., & Quinn, P. D. (2009). Development and validation of the Short Grit Scale (Grit-S). Journal of Personality Assessment, 91(2), 166-174. https://doi.org/10.1080/00223890802634290
31. Duke, N. K., Halvorsen, A.-L., Strachan, S. L., Kim, J., & Konstantopoulos, S. (2021). Putting PjBL to the test: The impact of project-based learning on second graders’ social studies and literacy learning and motivation in low-SES school settings. American Educational Research Journal, 58(1), 160-200. https://doi.org/10.3102/0002831220929638
32. Educating for American Democracy. (2021). Educating for American democracy: Excellence in history and civics for all learners [Report and roadmap]. iCivics. https://www.educatingforamericandemocracy.org/the-report/
33. Farrell, A. P. (Trans.). (1970). The Jesuit Ratio Studiorum of 1599: Translated into English, with an introduction and explanatory notes. Conference of Major Superiors of Jesuits. https://www.educatemagis.org/wp-content/uploads/documents/2019/09/ratio-studiorum-1599.pdf
34. Faure, E., Herrera, F., Kaddoura, A.-R., Lopes, H., Petrovsky, A. V., Rahnema, M., & Ward, F. C. (1972). Learning to be: The world of education today and tomorrow. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000001801
35. Fredricks, J. A., Blumenfeld, P. C., & Paris, A. H. (2004). School engagement: Potential of the concept, state of the evidence. Review of Educational Research, 74(1), 59-109. https://doi.org/10.3102/00346543074001059
36. Geger, B. T. (2012). What magis really means and why it matters. Jesuit Higher Education: A Journal, 1(2), 16-31. https://digitalcommons.lmu.edu/jhe/vol1/iss2/16/
37. Geger, B. T. (2014). Cura personalis: Some Ignatian inspirations. Jesuit Higher Education: A Journal, 3(2), 6-20. https://digitalcommons.lmu.edu/jhe/vol3/iss2/2/
38. Gorard, S., Siddiqui, N., & See, B. H. (2015). Philosophy for Children: Evaluation report and executive summary. Education Endowment Foundation. https://educationendowmentfoundation.org.uk/projects-and-evaluation/projects/philosophy-for-children
39. Grendler, P. F. (2014). Jesuit schools in Europe: A historiographical essay. Journal of Jesuit Studies, 1(1), 7-25. https://doi.org/10.1163/22141332-00101002
40. Hase, S., & Kenyon, C. (2000). From andragogy to heutagogy. ultiBASE In-Site. RMIT University. https://www.researchgate.net/publication/301339522_From_andragogy_to_heutagogy
41. Hirsch, E. D., Jr. (2016). Why knowledge matters: Rescuing our children from failed educational theories. Harvard Education Press. https://eric.ed.gov/?id=ED568821
42. Hmelo-Silver, C. E., Duncan, R. G., & Chinn, C. A. (2007). Scaffolding and achievement in problem-based and inquiry learning: A response to Kirschner, Sweller, and Clark (2006). Educational Psychologist, 42(2), 99-107. https://doi.org/10.1080/00461520701263368
43. Hrynowski, Z. (2024, August 20). K-12 schools struggle to engage Gen Z students. Gallup. https://news.gallup.com/poll/648896/schools-struggle-engage-gen-students.aspx
44. International Commission on the Apostolate of Jesuit Education. (1993). Ignatian pedagogy: A practical approach. International Center for Jesuit Education. https://jesuit.ie/wp-content/uploads/2019/09/Ignatian-Pedagogy.pdf
45. International Commission on the Futures of Education. (2021). Reimagining our futures together: A new social contract for education. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000379707
46. Jachimowicz, J. M., Wihler, A., Bailey, E. R., & Galinsky, A. D. (2018). Why grit requires perseverance and passion to positively predict performance. Proceedings of the National Academy of Sciences, 115(40), 9980-9985. https://doi.org/10.1073/pnas.1803561115
47. Jay, T., Willis, B., Thomas, P., Taylor, R., Moore, N., Burnett, C., Merchant, G., & Stevens, A. (2017). Dialogic Teaching: Evaluation report and executive summary. Education Endowment Foundation. https://educationendowmentfoundation.org.uk/projects-and-evaluation/projects/dialogic-teaching
48. Jay, T., Willis, B., Thomas, P., Taylor, R., Moore, N., Burnett, C., Merchant, G., & Stevens, A. (2019). Dialogic Teaching: Evaluation report addendum. Education Endowment Foundation. https://d2tic4wvo1iusb.cloudfront.net/production/documents/projects/Dialogic_Teaching_addendum_report__final_for_publication.pdf
49. Jubilee Centre for Character and Virtues. (2022). The Jubilee Centre framework for character education in schools (3rd ed.). University of Birmingham. https://www.jubileecentre.ac.uk/wp-content/uploads/2024/12/The-Jubilee-Centre-Framework-for-Character-Education-in-Schools.pdf
50. Kant, I. (1784). What is enlightenment? (M. C. Smith, Trans.). Columbia University. http://www.columbia.edu/acis/ets/CCREAD/etscc/kant.html
51. Kant, I. (1996). An answer to the question: What is enlightenment? In M. J. Gregor (Ed. & Trans.), Practical philosophy (pp. 11-22). Cambridge University Press. (Original work published 1784) https://doi.org/10.1017/CBO9780511813306
52. Kestin, G., Miller, K., Klales, A., Milbourne, T., & Ponti, G. (2025). AI tutoring outperforms in-class active learning: An RCT introducing a novel research-based design in an authentic educational setting. Scientific Reports, 15, Article 17458. https://doi.org/10.1038/s41598-025-97652-6
53. Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal guidance during instruction does not work: An analysis of the failure of constructivist, discovery, problem-based, experiential, and inquiry-based teaching. Educational Psychologist, 41(2), 75-86. https://doi.org/10.1207/s15326985ep4102_1
54. Kirschner, P. A., & van Merriënboer, J. J. G. (2013). Do learners really know best? Urban legends in education. Educational Psychologist, 48(3), 169-183. https://doi.org/10.1080/00461520.2013.804395
55. Knowles, M. S. (1980). The modern practice of adult education: From pedagogy to andragogy (Rev. ed.). Cambridge Adult Education. https://archive.org/details/modernpracticeof0000know_j2r4
56. Kohn, A. (1997). How not to teach values: A critical look at character education. Phi Delta Kappan, 78(6), 428-439. https://www.alfiekohn.org/article/teach-values/
57. Kolb, D. A. (1984). Experiential learning: Experience as the source of learning and development. Prentice-Hall. https://learningfromexperience.com/research-library/
58. Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X.-H., Beresnitzky, A. V., Braunstein, I., & Maes, P. (2025). Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant for essay writing task [Preprint]. arXiv. https://arxiv.org/abs/2506.08872
59. Krajcik, J., Schneider, B., Miller, E. A., Chen, I.-C., Bradford, L., Baker, Q., Bartz, K., Miller, C., Li, T., Codere, S., & Peek-Brown, D. (2023). Assessing the effect of project-based learning on science learning in elementary schools. American Educational Research Journal, 60(1), 70-102. https://doi.org/10.3102/00028312221129247
60. Kraut, R. (2026). Aristotle’s ethics. In E. N. Zalta & U. Nodelman (Eds.), The Stanford encyclopedia of philosophy. Stanford University. https://plato.stanford.edu/entries/aristotle-ethics/
61. Kristjánsson, K. (2015). Aristotelian character education. Routledge. https://doi.org/10.4324/9781315752747
62. Lei, H., Cui, Y., & Zhou, W. (2018). Relationships between student engagement and academic achievement: A meta-analysis. Social Behavior and Personality, 46(3), 517-528. https://doi.org/10.2224/sbp.7054
63. Lord, P., Dirie, A., Kettlewell, K., & Styles, B. (2021). Evaluation of Philosophy for Children: An effectiveness trial. Education Endowment Foundation. https://educationendowmentfoundation.org.uk/projects-and-evaluation/projects/philosophy-for-children-effectiveness-trial
64. MacArthur, C. A., & Cavalier, A. R. (2004). Dictation and speech recognition technology as test accommodations. Exceptional Children, 71(1), 43-58. https://doi.org/10.1177/001440290407100103
65. Mercer, N., & Littleton, K. (2007). Dialogue and the development of children’s thinking: A sociocultural approach. Routledge. https://oro.open.ac.uk/6519/
66. Michaels, S., O’Connor, C., & Resnick, L. B. (2008). Deliberative discourse idealized and realized: Accountable talk in the classroom and in civic life. Studies in Philosophy and Education, 27(4), 283-297. https://doi.org/10.1007/s11217-007-9071-1
67. Moore, D. L. (2007). Overcoming religious illiteracy: A cultural studies approach to the study of religion in secondary education. Palgrave Macmillan. https://doi.org/10.1057/9780230607002
68. Murphy, P. K., Wilkinson, I. A. G., Soter, A. O., Hennessey, M. N., & Alexander, J. F. (2009). Examining the effects of classroom discussion on students’ comprehension of text: A meta-analysis. Journal of Educational Psychology, 101(3), 740-764. https://doi.org/10.1037/a0015576
69. National Center for Education Statistics. (2023a). NAEP report card: 2022 NAEP civics assessment, grade 8: Achievement-level results. U.S. Department of Education. https://www.nationsreportcard.gov/civics/results/achievement/
70. National Center for Education Statistics. (2023b). NAEP report card: 2022 NAEP U.S. history assessment, grade 8: Achievement-level results. U.S. Department of Education. https://www.nationsreportcard.gov/ushistory/results/achievement/
71. National Institute of Child Health and Human Development. (2000). Report of the National Reading Panel: Teaching children to read: An evidence-based assessment of the scientific research literature on reading and its implications for reading instruction (NIH Publication No. 00-4769). U.S. Government Printing Office. https://www.nichd.nih.gov/publications/pubs/nrp/smallbook
72. Newfield, T. (2025). Heutagogy in education: Fostering self-directed learning for the 21st-century. Journal of Career and Technical Education, 40(1). https://eric.ed.gov/?id=EJ1479756
73. Nickow, A., Oreopoulos, P., & Quan, V. (2020). The impressive effects of tutoring on PreK-12 learning: A systematic review and meta-analysis of the experimental evidence (NBER Working Paper No. 27476). National Bureau of Economic Research. https://www.nber.org/papers/w27476
74. Nord, W. A. (1995). Religion and American education: Rethinking a national dilemma. University of North Carolina Press. https://uncpress.org/book/9780807844786/religion-and-american-education/
75. Nord, W. A., & Haynes, C. C. (1998). Taking religion seriously across the curriculum. ASCD. https://ascd.org/books/taking-religion-seriously-across-the-curriculum
76. Nussbaum, M. C. (2010). Not for profit: Why democracy needs the humanities. Princeton University Press. https://press.princeton.edu/books/paperback/9780691264394/not-for-profit
77. OECD. (2015). Students, computers and learning: Making the connection. OECD Publishing. https://doi.org/10.1787/9789264239555-en
78. O’Malley, J. W. (1993). The first Jesuits. Harvard University Press. https://www.hup.harvard.edu/books/9780674303133
79. Panta, R. (2025). Heutagogy: A comprehensive review of self-determined learning in contemporary education. Cureus, 17(8), Article e89731. https://doi.org/10.7759/cureus.89731
80. Pavur, C. (Trans.). (2005). The Ratio Studiorum: The official plan for Jesuit education. Institute of Jesuit Sources. https://jesuitsources.bc.edu/the-ratio-studiorum-the-official-plan-for-jesuit-education/
81. Recht, D. R., & Leslie, L. (1988). Effect of prior knowledge on good and poor readers’ memory of text. Journal of Educational Psychology, 80(1), 16-20. https://doi.org/10.1037/0022-0663.80.1.16
82. Reich, J. (2020). Failure to disrupt: Why technology alone can’t transform education. Harvard University Press. https://www.hup.harvard.edu/books/9780674278684
83. Reschly, A. L., & Christenson, S. L. (2012). Jingle, jangle, and conceptual haziness: Evolution and future directions of the engagement construct. In S. L. Christenson, A. L. Reschly, & C. Wylie (Eds.), Handbook of research on student engagement (pp. 3-19). Springer. https://doi.org/10.1007/978-1-4614-2018-7_1
84. Resnick, L. B., Asterhan, C. S. C., & Clarke, S. N. (Eds.). (2015). Socializing intelligence through academic talk and dialogue. American Educational Research Association. https://doi.org/10.3102/978-0-935302-43-1
85. Risko, E. F., & Gilbert, S. J. (2016). Cognitive offloading. Trends in Cognitive Sciences, 20(9), 676-688. https://doi.org/10.1016/j.tics.2016.07.002
86. Saavedra, A. R., Liu, Y., Haderlein, S. K., Rapaport, A., Garland, M., Hoepfner, D., Lock Morgan, K., & Hu, A. (2021). Knowledge in Action efficacy study over two years. USC Dornsife Center for Economic and Social Research. https://cesr.usc.edu/sites/default/files/Knowledge%20in%20Action%20Efficacy%20Study_18feb2021_final.pdf
87. Saavedra, A. R., Lock Morgan, K., Liu, Y., Garland, M. W., Rapaport, A., Hu, A., Hoepfner, D., & Haderlein, S. K. (2022). The impact of project-based learning on AP exam performance. Educational Evaluation and Policy Analysis, 44(4), 638-666. https://doi.org/10.3102/01623737221084355
88. Schwickerath, R. (1903). Jesuit education: Its history and principles viewed in the light of modern educational problems. B. Herder. https://en.wikisource.org/wiki/Jesuit_Education/Chapter_4
89. Schwickerath, R. (1911). Ratio Studiorum. In The Catholic Encyclopedia (Vol. 12). Robert Appleton Company. https://www.newadvent.org/cathen/12654a.htm
90. Societas Iesu. (1606). Ratio atque institutio studiorum Societatis Iesu. In Collegio Romano eiusdem Societatis. (Original work published 1599) https://archive.org/details/bub_gb_6mcuvbXlgXEC
91. Suissa, J. (2015). Character education and the disappearance of the political. Ethics and Education, 10(1), 105-117. https://doi.org/10.1080/17449642.2014.998030
92. UNESCO. (2023). Global education monitoring report 2023: Technology in education. A tool on whose terms? UNESCO. https://doi.org/10.54676/UZQV8501
93. Willingham, D. T. (2007). Critical thinking: Why is it so hard to teach? American Educator, 31(2), 8-19. https://www.aft.org/sites/default/files/periodicals/Crit_Thinking.pdf
94. Wong, Z. Y., & Liem, G. A. D. (2022). Student engagement: Current state of the construct, conceptual refinement, and future research directions. Educational Psychology Review, 34(1), 107-138. https://doi.org/10.1007/s10648-021-09628-3
95. Wood, S. G., Moxley, J. H., Tighe, E. L., & Wagner, R. K. (2018). Does use of text-to-speech and related read-aloud tools improve reading comprehension for students with reading disabilities? A meta-analysis. Journal of Learning Disabilities, 51(1), 73-84. https://doi.org/10.1177/0022219416688170
96. Wyse, D., & Bradbury, A. (2022). Reading wars or reading reconciliation? A critical examination of robust research evidence, curriculum policy and teachers’ practices for teaching phonics and reading. Review of Education, 10(1), Article e3314. https://doi.org/10.1002/rev3.3314