Paper 4 of Five
Abstract
This paper names a mechanism by which an indefensible practice becomes defensible without being defended: it is relocated into a register where the governing text is not consulted, and where its non-consultation reads not as evasion but as ordinary professional propriety. I call this domain quarantine. The claim is that a question can be moved from the pulpit to the bench, from the bench to the counting-house, from theology to political economy, and that at each move the standard that would have condemned it becomes, by the receiving register’s own rules, out of place. Three detection tests are given: register vocabulary shift, author overlap across venues, and the missing-defense count. The paper argues that this is the hardest signature to see from inside, because from inside the quarantine feels like ordinary division of labor — and because division of labor is genuinely valuable, which is what makes the mechanism both effective and difficult to indict. Antebellum American religious and legal argument supplies the demonstration; the paper also argues that domain quarantine is the strongest candidate in the series for transfer to present-day cases, and treats the danger of that transfer at length rather than in a closing caveat. Falsification constraint: if the register distribution of a governing text’s citation matches the register distribution of comparable texts on questions where no quarantine is alleged, the category has no referent.
1. The Move
The 1850–1851 rendition sermons are Paper 1’s paradigm case, and they contain a formula that appears often enough to be worth taking seriously as a structural device rather than a rhetorical tic. Some version of: the question of the law’s justice belongs to the legislator and the jurist, not to this pulpit; our office is to teach obedience to the ordinance lawfully constituted.
Turn to the bench. Judges of the period — including judges whose private antislavery convictions are documented — held with striking regularity that the moral status of the statute was not before them. Their office was to construe what the legislature had enacted. Conscience belonged elsewhere: to the pulpit, to the legislature, to the individual citizen.
Turn to the legislature. There, the question was compromise, union, the terms on which a federation of unlike parts could hold together. A prudential question, and honorably so. The scriptural question was for the divines.
At each station a competent professional performed his office correctly, and at each station the standard that would have settled the matter was, by the local rules of that station, not the business of the person present. The minister deferred to law. The judge deferred to legislation. The legislator deferred to prudence and to the divines. Nobody was silent. Nobody refused an argument he had been handed. The governing text was simply never in the room where the decision was made, because at every location where the decision was being made it belonged to some other room.
That is domain quarantine. The practice has been relocated to a register whose governing standards do not include the text that would condemn it, and the relocation is invisible because each individual deferral is a normal and often admirable professional restraint.
2. The Mechanism Stated
Domain quarantine is the relocation of a contested practice into a discursive register in which the standard that would govern it is, by the receiving register’s own conventions, out of scope.
Four features:
(a) A governing standard exists and is acknowledged. All parties concede that the standard, in its own domain, is binding. This is what distinguishes quarantine from simple secularization. Nobody is denying that Scripture governs; they are locating the question elsewhere.
(b) The practice is relocated, not the standard. The standard stays where it is and retains full authority there. What moves is the question. This is the reverse of the usual account of secularization, in which a standard loses authority; here it keeps authority and loses jurisdiction.
(c) The receiving register has its own legitimate exclusion rule. Courts do not decide on Scripture. Political economy does not reason from commandments. Pulpits do not adjudicate statutory construction. Each of these exclusions is defensible on its own terms, and the defense is not a pretext — it is the register’s actual constitutive rule.
(d) No single actor performs the relocation. The move is distributed. Each participant hands the question along, and each hand-off is professionally correct. There is no point at which someone decides to put the question beyond reach.
Feature (d) is the whole difficulty. Functional subtraction (Paper 1) attributes a selection to a citer. Unexamined title (Paper 2) attributes a non-litigation to a dispute. Domain quarantine attributes nothing to anyone. It is a property of a system of registers, and it can be produced by participants each of whom is behaving well.
3. Why It Is Hardest to See From Inside
Division of labor is one of the great achievements of institutional life. A judge who decided cases on his private theology would be a bad judge and a menace. A minister who ruled on statutory construction from the pulpit would be exceeding his competence and his office. A political economist who reasoned from commandments would produce bad economics. The exclusions are not corruptions of these practices; they are what make them practices.
This means the quarantine feels, from every station inside it, exactly like professional integrity. The minister who says the justice of the statute is not his to determine is not making an excuse. He is doing what a man of modest office should do — declining to rule on a matter outside his competence. He may even feel the restraint as a cost.
And in most cases he is right. That is the difficulty stated at its sharpest: the behavior that constitutes the quarantine is, in the general run of cases, correct behavior. Any detection method must therefore separate correct restraint from quarantine, and cannot do so by pointing at the restraint itself, because the restraint is identical in both.
The separation has to come from the pattern across registers. A single deferral is invisible. A closed loop of deferrals — where every register hands the question to another and the question comes to rest nowhere — is not, provided one is looking at the whole loop. But looking at the whole loop requires standing outside all the registers at once, which is precisely what a participant cannot do and a historian can.
This is also why the paper carries the largest risk of the series. A method that detects a wrong constituted by no one’s wrongdoing is a method that can convict anyone. §7 treats this at the length it deserves.
4. Three Detection Tests
4.1 T1 — Register vocabulary shift
The same practice is described in different registers by non-overlapping vocabularies, and the vocabularies are not translations of one another but replacements.
In the pulpit: servants, masters, households, obedience, duty, the relation. In the courts: property, chattels, conveyance, warranty, title, comity, rendition. In political economy: labor, capital, productivity, the staple, the system.
A vocabulary shift is innocent where the registers describe genuinely different aspects of a thing, and terms translate between them. A shift is a quarantine signature where the register’s vocabulary makes the governing standard unstatable. The test is not “do the words differ” but “can the command be put into this register’s terms without ceasing to be a command.”
Operationally: take the governing text, translate it into each register’s working vocabulary, and ask whether the translation still says anything. Give unto your servants that which is just and equal enters legal register as a question about the standard of care owed to chattel property, which is a different proposition entirely — it has become a question about the treatment of a thing rather than an obligation to a person before God. It enters economic register as a question about the efficient level of maintenance expenditure. Neither translation preserves the command; both dissolve it into the receiving register’s terms.
The signature fires when the translation dissolves rather than transfers.
4.2 T2 — Author overlap across venues
The same individuals write in both registers, and their treatment of the question differs by venue in ways not explained by the venue’s subject matter.
This is the sharpest of the three tests, because it holds the person constant. A man who argues in a denominational quarterly that the household codes bind masters absolutely, and in a legal or economic paper treats the same relation entirely in terms of property and productivity without the obligation appearing, has performed the relocation in his own hands. His two selves are the two registers, and the gap between them is the quarantine made visible in a single biography.
The period supplies unusual material for this test because the professional classes overlapped heavily. Ministers sat in legislatures. Judges were elders and vestrymen. Editors of denominational quarterlies wrote on political economy. College presidents wrote theology and moral philosophy and addressed commercial conventions.
Operationally: build an author register of individuals with documented output in two or more registers within the window; for each, code the treatment of the governing text in each register using the Paper 3 C1–C3 scheme; and report the within-author, between-register difference. Within-author comparison controls for belief, competence, era, and personal position, all of which the between-author comparison leaves open. This is the same logic as Paper 3’s within-passage design applied to persons instead of verses, and it inherits both the strength and the small-n problem.
4.3 T3 — The missing-defense count
In a quarantined register, the practice generates no defense, because no attack is intelligible in that register’s terms.
The count is of defenses, not of mentions. A practice under moral scrutiny generates apologetics: justifications, distinctions, arguments that it is not what it appears. A practice in quarantine generates none — it generates administration. Cases about warranty on a defective sale. Analyses of the staple’s productivity. Rules for the disposition of an estate. The practice is everywhere as a premise and nowhere as a proposition.
The signature is a high mention rate with a near-zero defense rate in a register where the practice is heavily present. Where a register both mentions and defends, the question is live there and no quarantine obtains. Where a register neither mentions nor defends, the register is simply not about the practice. Only the mention-without-defense combination is diagnostic.
This test is the most easily automated and the most easily misread, because “defense” requires coding a communicative function rather than a topic. Coding scheme and failure modes at note 4.
5. Relation to the Rest of the Series
The four papers describe four different objects and the boundaries need stating, because they overlap in the demonstration corpus and a reader could reasonably take them for one claim in four costumes.
| Unit | Question | |
| Paper 1 | The text | Was this verse brought forward where it bore? |
| Paper 2 | The question | Was this topic a site of dispute? |
| Paper 3 | The passage half | Was the citer-binding half cited alongside the other? |
| Paper 4 | The register | Was the standard in scope where the decision was made? |
These come apart. A text can be subtracted in a register where it is fully in scope (Paper 1 without Paper 4). A question can go unlitigated with no governing text existing (Paper 2 without Paper 1). A register can exclude a standard on a question nobody contests, which is quarantine’s form without its object and is not a finding at all.
Where they coincide — as they do repeatedly in the demonstration corpus — the case enters the joint register once. The series does not treat coincidence as independent confirmation, and a reader should be suspicious of any presentation that seems to.
One live migration. Paper 3’s A5 condition provides that if the master-binding halves prove well represented in devotional and pastoral registers and thin only in controversy, that finding migrates here. The attempted register test on Paper 3 was not completed for infrastructure reasons, and its single scoping observation was mixed: inward-binding material present in the pastoral register, and half-verse warrant also present there. The refinement that came out of that attempt — present versus brought to bear — is directly this paper’s distinction, since bringing a standard to bear is a question about scope rather than about presence. If the split hypothesis H4a/H4b resolves as H4a-fails-H4b-holds, the finding belongs to both papers and should be reported as one.
6. Falsification Constraint
If the register distribution of the governing text’s citation, in the demonstration corpus, matches the register distribution of comparable texts on questions where no quarantine is alleged, then domain quarantine has no referent and this paper should be withdrawn.
Four subsidiary constraints, each independently sufficient:
F1 — Translation parity. If T1’s dissolution test, applied blind to governing texts on non-quarantined questions, dissolves them at comparable rates, then dissolution is a general property of cross-register translation and carries no weight. This is the most likely of the four to fire, because much theological language translates poorly into technical registers for reasons having nothing to do with evasion.
F2 — Author consistency. If within-author, between-register differences (T2) are no larger for the demonstration question than for control questions handled by the same authors, then the difference is a register effect on all topics and not a quarantine.
F3 — Defense-rate parity. If mention-without-defense rates in the legal and economic registers are comparable for the demonstration practice and for uncontroversial practices of similar prominence, then T3 is measuring the ordinary character of technical writing.
F4 — Loop non-closure. If the deferral chain does not close — if tracing the hand-offs leads to a register where the question is in scope and is decided on the merits — then the practice was not quarantined but adjudicated, and the paper has mistaken a division of labor for an enclosure. This is the strongest test and the one I would run first.
All four are reported whichever way they fall. F4 in particular should be run before the others, because a non-closing loop makes the remaining tests pointless.
7. Steelmanning: The Strongest Case Against
What follows is the best case I can build against this paper. Two of these I regard as damaging and one I cannot answer at all.
7.1 This is just secularization, and it is a good thing
The objection. What the paper describes is the differentiation of modern institutional life: law becomes autonomous from theology, economics from moral philosophy, the professions from the church. This process is well documented, is the subject of a large literature, and is on balance a benefit. It gave us courts that do not burn heretics and economies not run by clergy. Calling it “quarantine” — a word borrowed from disease containment — smuggles in a verdict that the analysis has not earned.
The strongest form. Press further: the paper’s own §3 concedes that the exclusions are constitutive rather than pretextual. If a court’s refusal to decide on Scripture is what makes it a court, then complaining that the court did not decide on Scripture is complaining that a court is a court. The paper wants the benefits of differentiation and the right to indict it selectively, and has no principle for when the indictment applies beyond the investigator’s prior conviction that the outcome was wrong.
Response. The objection is right that differentiation is real and beneficial and that the paper’s vocabulary is loaded. I take the vocabulary point and have no better word; readers should substitute the neutral gloss “register relocation” wherever the connotation grates.
On the principle: F4 supplies it, and it is the paper’s only answer. Differentiation is benign when the deferral chain closes on a competent forum — when the question the court declines is one the legislature or the pulpit or the electorate actually takes up and decides on the merits. Differentiation produces quarantine only when the chain closes on itself, so that every forum defers and none decides. That is a structural property, checkable without any judgment about the outcome’s rightness, and it is what F4 tests.
Whether this answer holds depends entirely on whether F4 can be operationalized cleanly enough to be run by someone who disagrees with me about slavery. I think it can. I have not shown it.
7.2 The chain did close — the objection the paper cannot dismiss
The objection. This is the serious one. The deferral chain in the demonstration case did close, repeatedly and publicly.
The question went to the legislature and the legislature decided it — in 1808, in 1820, in 1850, in 1854. It went to the courts and the courts decided it, up to and including the Supreme Court in 1857. It went to the churches and the churches decided it: the Methodists split in 1844, the Baptists in 1845, the Presbyterians before and after. A denominational schism is not a deferral. It is the most costly form of deciding a question that a church has available.
And the question went to the electorate, which decided it in 1860, and then to the field, which decided it between 1861 and 1865. Whatever else may be said of the antebellum republic, it did not fail to confront this question. It confronted it more directly and at greater cost than any other question in its history.
The strongest form. So F4 fires against the paper on the demonstration case, and F4 is the paper’s own designated primary test. The category may be coherent in the abstract, but the case chosen to demonstrate it is the worst possible case, because it is the historical instance in which deferral most conspicuously failed.
Response. I do not think I can answer this, and I want to be precise about where the failure sits rather than negotiating it down.
What I can say: the chain’s closing was late and was forced from outside the registers rather than by any of them. Every closure the objection names came from a forum the quarantine had not anticipated — a mass electorate, a schism driven substantially by lay pressure, and finally arms. None came from a professional register conceding jurisdiction. On that reading the quarantine held within the register system and was broken by something external to it, which is compatible with the category and even a point in its favor: the measure of an enclosure is what it takes to breach it.
But I notice that this response has the shape of an unfalsifiable move. If the chain doesn’t close, that confirms quarantine; if it closes, I say it closed from outside and that also confirms quarantine. That is exactly the pattern the reproducibility literature identifies as a hypothesis that has stopped being a hypothesis. A reader is entitled to conclude that F4 has fired and that the demonstration case fails.
Two honest consequences. First, F4 must be restated to specify in advance what counts as closure from inside versus outside the register system, and that specification must be written before the tracing is done, or F4 is not a test. Second, and more seriously: if the demonstration case fails F4, the right response is to say so and look for a case where the chain genuinely does not close, rather than to defend this one. I do not currently have such a case in hand. That is a gap in the paper and not a rhetorical concession.
7.3 T2 proves too much about ordinary professional life
The objection. Every professional writes differently in different venues, and none of us regards this as evidence of anything. A physician’s clinical notes differ from her conversation with a patient’s family; a lawyer’s brief differs from his advice; a scholar’s journal article differs from her public writing. The differences are large, systematic, and entirely proper. T2 will detect them everywhere and will therefore convict everyone.
Response. F2 is the answer and it is a real test: the comparison is not between registers but between the size of the between-register gap for the demonstration question and for control questions handled by the same authors in the same venues. A physician whose notes differ from her family conversations on every topic equally has shown nothing. One whose gap is ten times larger on one specific diagnosis has shown something.
The residual problem the objection correctly identifies is that control questions must be matched on stakes, not merely on topic, and I do not have a good way to match on stakes. High-stakes questions may generate larger register gaps for reasons entirely unrelated to quarantine — caution, liability, the desire not to be quoted. This is unresolved and F2 should be reported with it noted.
7.4 The method is a general-purpose accusation engine
The objection. This is the concern the paper’s own §3 raises and it deserves the strongest form available. Domain quarantine detects a wrong constituted by nobody’s wrongdoing. It requires no evidence of intent, no showing of bad faith, no identifiable actor. It runs on structural features that are present in every differentiated institution. And the paper explicitly nominates it as the series’ best candidate for transfer to present-day cases.
That combination is dangerous in a specific way: it lets a user indict an institution without having to establish that anyone in it did anything wrong, and it supplies a scholarly apparatus — three tests, four falsification constraints, a coding scheme — that makes the indictment look measured. The apparatus is the danger, not the thesis. A bare accusation can be evaluated. An accusation with a method behind it borrows the method’s credibility.
Response. I accept the framing and do not think the mitigations are adequate.
What mitigations exist. F4 requires demonstrating a closed loop, which is a positive empirical showing and a demanding one — the investigator must trace every deferral and show the question comes to rest nowhere. F1 through F3 all require matched controls, which are expensive and which a motivated user will be tempted to skip; skipping them is detectable, and this paper says so in advance so that the skip can be named as a defect rather than passing as a shortcut. And the method’s output is a structural description rather than a verdict on persons, which is a real limit on what it licenses even when it fires.
What those mitigations do not do. They do not prevent misuse; they make misuse identifiable to a careful reader, which is a weaker thing. They rely on the user’s honesty about the controls, which is exactly what a motivated user will not supply. And they do nothing about the deeper problem the objection names, which is that scholarly apparatus lends credibility independent of whether the apparatus was correctly applied.
I have considered whether the right response is to withhold the transfer claim — to present the mechanism historically and decline to nominate it for present-day use. I have decided against, on the grounds that a mechanism which only ever operates on the safely dead is not a methodological contribution. But I hold that judgment loosely, and a reader who concludes the tool should not be published in transferable form has reached a position I can respect. The series’ own commitment is that a failed paper is published as failed; a paper judged too dangerous to publish would be a different outcome, and I have not established a procedure for reaching it.
7.5 The registers were not separate
The objection. The paper’s picture requires distinct registers with distinct rules. But antebellum American discourse was not compartmentalized in this way. Political speeches quoted Scripture constantly. Judicial opinions of the period made moral arguments openly. Denominational quarterlies ran political economy. Congressional debate on the 1850 measures was saturated with biblical reference on both sides. The “registers” are an anachronistic imposition of later professional differentiation on a period that had much less of it.
Response. This is a fair historical objection and it cuts against T1 in particular. The period’s registers were more permeable than my framing suggests, and if they were permeable then the governing text was not excluded by rule but simply not raised — which returns the phenomenon to Paper 1 and leaves this paper with no distinct object.
The partial answer is that permeability is itself measurable, and should be measured before the quarantine claim is made: what proportion of judicial opinions in the period contain scriptural reference at all? What proportion of political-economic writing? If the proportion is substantial, the exclusion rule was not operating and T1 should be withdrawn for this corpus. This measurement is prerequisite and is not currently in hand. T2 and T3 survive the objection better than T1, since both turn on differential treatment rather than on categorical exclusion.
7.6 Quarantine is a description of relief, not evasion
The objection. Consider the minister who believed the institution wrong, held a congregation containing men who held people, and had a family to feed. The doctrine that the justice of the statute was not his to determine was not a device for avoiding a conclusion. It was what made it possible for him to keep his post and go on preaching. What the paper calls a quarantine wall may have been, for many participants, a load-bearing structure that let them continue to function at all.
Response. I think this is right and it does not defeat the category, though it changes what the category is a description of. The mechanism can be simultaneously a producer of negative space and a mercy to the people inside it. Structures that let people function under pressure and structures that prevent questions from being decided are not different structures; they are the same structure described from two positions.
What follows methodologically is that the paper must not read as an indictment of the individuals, and I am not confident the current draft achieves that. The vocabulary at §1 and §4 leans toward accusation in places where the analysis does not support it, and that should be corrected in revision rather than defended.
8. Scope and What Is Not Claimed
The paper does not claim that differentiation of registers is a defect. It is an achievement, and the paper’s own test (F4) is built to return “benign” for the ordinary case.
It does not claim intent, awareness, or bad faith on the part of any participant. §2(d) is explicit that no single actor performs the relocation, and §7.6 concedes that the structure may have functioned as relief.
It does not claim that the demonstration case has been established. §7.2 concedes that F4 appears to fire against it and that I have no replacement case in hand. A reader who takes this paper as an argument for the mechanism’s coherence and not as a demonstration of its instantiation has read it correctly.
It does not claim that the transfer to present-day cases is safe. §7.4 states the danger and reports that the available mitigations are inadequate to it.
Notes
- On the word “quarantine.” Borrowed from containment and carrying an unearned verdict, per §7.1. The neutral gloss is “register relocation.” I have not found a term that names the phenomenon without either prejudging it or losing it, and if one exists this paper should adopt it.
- F4 specification is prerequisite. Following §7.2, the criteria distinguishing closure-from-inside from closure-from-outside must be written and fixed before any deferral chain is traced. Closure by mass electorate, by armed conflict, by lay revolt against clerical leadership, and by schism each require an advance ruling. Tracing first and classifying after is not a test.
- Register permeability measurement is prerequisite. Following §7.5, the base rate of scriptural citation in judicial opinions, legislative debate, and political-economic writing in the window must be established before T1 is applied. If permeability is high, T1 is withdrawn for this corpus.
- Coding “defense” for T3. A defense is an argumentative move asserting that the practice is licit, just, or required, offered against an actual or anticipated denial. Administrative treatment (rules for conveyance, valuation, disposition) is not a defense even where the practice’s licitness is presupposed throughout. The commonest coding failure is scoring presupposition as defense; the second commonest is scoring a defense of an incident of the practice as a defense of the practice. Both are addressed in the training set and both should be reported in the reliability tables separately.
- Author register for T2. Inclusion requires documented output in two or more registers within the window under the same name, with authorship attribution independently verifiable. Anonymous and pseudonymous periodical writing is excluded, which will bias the register toward prominent men and away from the working clergy and bar. The bias is reported and not corrected, since correcting it would require attribution guesses.
- Reliability. All coding in this paper uses the Paper 3 apparatus without modification: three coders, two-of-three concordance, Krippendorff’s alpha floor of 0.67, indeterminate ceiling of 35%, no recoding to reach threshold. T3’s defense code is expected to be the hardest and should be reported separately rather than pooled.
- Relation to Paper 3’s migration. Per §5, if H4a fails and H4b holds, the finding is reported jointly and once. The present versus brought to bear distinction developed in that attempt should be treated as this paper’s operational core, since scope rather than presence is what quarantine governs.
- Relation to Paper 5. T1, T2, and T3 each require a matched control corpus and the control constructions live in Paper 5. Nothing here is established until they exist.
- Scripture. Quotations follow the Authorized Version, for the reason given in Papers 1 through 3.
- On what remains undone. Three items are prerequisite and none is in hand: the F4 advance specification (note 2), the permeability base rate (note 3), and a replacement demonstration case should F4 fire against the present one (§7.2). The paper should not be presented as established until the first two are complete and the third is either supplied or the failure is reported as a failure.
References
Abbott, A. (1988). The system of professions: An essay on the division of expert labor. University of Chicago Press.
Barnes, A. (1846). An inquiry into the Scriptural views of slavery. Parry & McMillan.
Bourne, G. (1845). A condensed anti-slavery Bible argument. S. W. Benedict.
Bowker, G. C., & Star, S. L. (1999). Sorting things out: Classification and its consequences. MIT Press.
Cover, R. M. (1975). Justice accused: Antislavery and the judicial process. Yale University Press.
Cover, R. M. (1983). Nomos and narrative. Harvard Law Review, 97(1), 4–68.
Dabney, R. L. (1867). A defence of Virginia, and through her, of the South. E. J. Hale & Son.
Douglas, M. (1986). How institutions think. Syracuse University Press.
Elliott, E. N. (Ed.). (1860). Cotton is king, and pro-slavery arguments. Pritchard, Abbott & Loomis.
Fehrenbacher, D. E. (1978). The Dred Scott case: Its significance in American law and politics. Oxford University Press.
Fehrenbacher, D. E. (2001). The slaveholding republic: An account of the United States government’s relations to slavery. Oxford University Press.
Finkelman, P. (1981). An imperfect union: Slavery, federalism, and comity. University of North Carolina Press.
Fox-Genovese, E., & Genovese, E. D. (2005). The mind of the master class: History and faith in the Southern slaveholders’ worldview. Cambridge University Press.
Freidson, E. (2001). Professionalism: The third logic. University of Chicago Press.
Genovese, E. D. (1998). A consuming fire: The fall of the Confederacy in the mind of the white Christian South. University of Georgia Press.
Goffman, E. (1974). Frame analysis: An essay on the organization of experience. Harvard University Press.
Goodell, W. (1852). The American slave code in theory and practice. American and Foreign Anti-Slavery Society.
Hayes, A. F., & Krippendorff, K. (2007). Answering the call for a standard reliability measure for coding data. Communication Methods and Measures, 1(1), 77–89.
Haynes, S. R. (2002). Noah’s curse: The biblical justification of American slavery. Oxford University Press.
Holifield, E. B. (2003). Theology in America: Christian thought from the age of the Puritans to the Civil War. Yale University Press.
Hopkins, J. H. (1864). A Scriptural, ecclesiastical, and historical view of slavery. W. I. Pooley.
Jones, C. C. (1842). The religious instruction of the Negroes in the United States. Thomas Purse.
Krippendorff, K. (2018). Content analysis: An introduction to its methodology (4th ed.). SAGE.
Luhmann, N. (1982). The differentiation of society (S. Holmes & C. Larmore, Trans.). Columbia University Press.
MacIntyre, A. (1981). After virtue: A study in moral theory. University of Notre Dame Press.
Mathews, D. G. (1977). Religion in the Old South. University of Chicago Press.
Morris, T. D. (1996). Southern slavery and the law, 1619–1860. University of North Carolina Press.
Noll, M. A. (2006). The Civil War as a theological crisis. University of North Carolina Press.
Nosek, B. A., Ebersole, C. R., DeHaven, A. C., & Mellor, D. T. (2018). The preregistration revolution. Proceedings of the National Academy of Sciences, 115(11), 2600–2606.
Proctor, R. N., & Schiebinger, L. (Eds.). (2008). Agnotology: The making and unmaking of ignorance. Stanford University Press.
Raboteau, A. J. (1978). Slave religion: The “invisible institution” in the antebellum South. Oxford University Press.
Ross, F. A. (1857). Slavery ordained of God. J. B. Lippincott.
Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2011). False-positive psychology: Undisclosed flexibility in data collection and analysis allows presenting anything as significant. Psychological Science, 22(11), 1359–1366.
Snay, M. (1993). Gospel of disunion: Religion and separatism in the antebellum South. Cambridge University Press.
Stringfellow, T. (1856). Scriptural and statistical views in favor of slavery. J. W. Randolph.
Swartley, W. M. (1983). Slavery, Sabbath, war, and women: Case issues in biblical interpretation. Herald Press.
Taylor, C. (2007). A secular age. Harvard University Press.
Thornwell, J. H. (1850). The rights and the duties of masters. Walker & James.
Tise, L. E. (1987). Proslavery: A history of the defense of slavery in America, 1701–1840. University of Georgia Press.
Trouillot, M.-R. (1995). Silencing the past: Power and the production of history. Beacon Press.
Tushnet, M. V. (1981). The American law of slavery, 1810–1860: Considerations of humanity and interest. Princeton University Press.
Vaughan, D. (1996). The Challenger launch decision: Risky technology, culture, and deviance at NASA. University of Chicago Press.
Weber, M. (1978). Economy and society (G. Roth & C. Wittich, Eds.). University of California Press. (Original work published 1922)
Weld, T. D. (1838). The Bible against slavery. American Anti-Slavery Society.
Wiecek, W. M. (1977). The sources of antislavery constitutionalism in America, 1760–1848. Cornell University Press.
