← The catalogue
SSRN · Methodology · 2026

Iterative Adversarial Inquiry: The Whetstone Method

Olorunfemi Obasa, MBA · Dr. Precious Obasa (née Olusegun)

First published on SSRN. The full text is reproduced here and the original stays linked.Read the original on SSRN

Reproduced in full from the SSRN working paper (April 2026). The correspondence address printed in the original is left out. The original stays linked so the two copies remain connected.

Extrafemi Research Paper No. 26-002

In plain words

The Whetstone method sharpens an argument by attacking it. A counter-position is written at full strength and the draft has to survive it. The aim is a genuine new idea at the end, not a summary of what is already known.

Published
27 April 2026
Reading time
35 min

Working paper, SSRN Working Paper Series, April 2026. This paper has not undergone peer review. Comments welcome.

Abstract

This paper introduces Iterative Adversarial Inquiry (IAI), colloquially known as the Whetstone Method, a framework for generating original theoretical contributions through structured adversarial debate. The dominant model of social science research, the hypothetico-deductive method, begins with a hypothesis and tests it against data. IAI begins differently: with a provocation derived from a researcher's existing, post-experiential bias, which is then subjected to adversarial pressure in two sequential debate stages: a logical debate that tests internal coherence, and a literature debate that tests novelty and positioning. The framework is cyclical rather than linear. The output of each iteration becomes the starting point of the next, producing a progressively refined body of knowledge in which each element carries a documented record of the challenges it has survived. We argue that this method is epistemologically superior to hypothesis-first research for one specific class of problem: generating the theoretical category that does not yet exist. We demonstrate this claim using a detailed case study: the development of the Aspiration Horizon framework, which emerged through precisely this process. We trace the method's philosophical antecedents in Peirce's abductive inquiry and self-correcting epistemology, distinguish it from Grounded Theory and the Delphi method, and specify its quality criteria, sparring partner requirements, and documentation standards. We further examine the method's applicability across disciplines, including management, policy, psychology, economics, and artificial intelligence research, and argue that the emergence of AI as a thinking interlocutor makes IAI available at scale for the first time in the history of academic research.

Keywords: iterative adversarial inquiry, Whetstone Method, research methodology, theoretical generation, abductive reasoning, adversarial debate, human-AI collaboration, epistemology

Journal of Economic Literature (JEL) Classification: B40, C90, Z13

1. Introduction

Research methodology is rarely where the interesting arguments happen. Most researchers learn a method, apply it, and move on to the substantive question. The method is infrastructure, necessary, but not itself the contribution. This paper takes a different view. It argues that the method by which a theoretical framework is generated is not merely infrastructure. In certain classes of problems, it is the most important variable in determining whether the framework produced is original, defensible, and genuinely additive to existing knowledge.

The dominant model of social science research: the hypothetico-deductive (HD) method, rooted in Popper's (1959) falsificationism, begins with a hypothesis derived from prior theory and tests it against data. This model has been enormously productive. It has produced decades of cumulative knowledge across economics, sociology, psychology, and demography. But it has a structural limitation that is rarely acknowledged directly: it can only generate hypotheses from within existing theoretical categories. You cannot pre-register a hypothesis about something you do not yet have a name for. You cannot operationalise a construct that has not yet been defined. The hypothetico-deductive model is, by design, a machine for testing what is already imaginable. It is not a machine for generating what is not yet imaginable.

This paper introduces a method designed for precisely that second problem: generating the theoretical category that did not previously exist. We call it Iterative Adversarial Inquiry (IAI). The working name, which we believe captures the method's essential character more memorably, is the “Whetstone Method”. The name comes from the method's central analogy: the debate does not protect the framework from challenge the way armour protects a soldier from attack. It sharpens the framework against challenge, the way a whetstone sharpens a blade. The friction is the point. Survival under friction is evidence of quality.

We demonstrate the method through a detailed case study: the development of the Aspiration Horizon framework (Obasa & Obasa, 2026), which proposes that fertility decline in wealthy societies is driven primarily by the expansion of imaginable possible selves that development produces, rather than by fear, economic insecurity, or cultural value change. That framework did not begin with a literature review. It began with a provocation, was sharpened through a structured adversarial debate, survived a formal stress-test of five substantive challenges, and emerged with greater precision and depth than could have been produced through either solo reflection or conventional hypothesis-driven research. The case study is not incidental to the methodology paper. It demonstrates the method. The framework is the evidence that the method works.

The remainder of this paper proceeds as follows. Section 2 defines the key terms. Section 3 reviews the methodological landscape and locates IAI relative to existing methods. Section 4 develops the IAI framework in full, specifying each stage, its quality criteria, and its procedural requirements. Section 5 presents the Aspiration Horizon case study as a worked demonstration. Section 6 examines cross-disciplinary applications. Section 7 addresses the role of artificial intelligence (AI) as a sparring partner. Section 8 presents the stress-test record: the five challenges raised against the framework and the responses that survived. Section 9 concludes.

2. Key Terms and Definitions

Precision of language is not incidental to this paper. IAI is a framework for generating theoretical categories, and the argument requires that its own central terms be defined with the rigour it asks of the frameworks produced through it. Every term used in a technical sense is defined here at first use and consistently thereafter.

Iterative Adversarial Inquiry (IAI)

A cyclical research methodology in which theoretical frameworks are generated through structured adversarial debate, refined through sequential logical and literature stress-tests, and progressively strengthened over multiple iterations. Also referred to throughout this paper as the Whetstone Method. Formally named the Obasa Method of Iterative Adversarial Inquiry.

The Whetstone Method

The working name for IAI. The whetstone analogy specifies the method's relationship to challenge: a whetstone sharpens a blade rather than deflecting a blow. The adversarial debate does not protect the framework from objection. It sharpens the framework through objection. A framework that has survived genuine adversarial pressure is stronger because it has been sharpened, not because it has been insulated.

Provocation

The explicit, deliberately overdrawn statement of the researcher's post-experiential bias that initiates the first debate cycle. A provocation is not a hypothesis. A hypothesis is tentative and qualified. A provocation is stated with maximum clarity and force, precisely so that the strongest available objections can be brought against it. The overstatement is methodologically functional: it generates more productive friction than a carefully hedged claim.

Post-experiential bias

The prior orientation toward a phenomenon that a researcher carries as a result of prior encounter, observation, professional experience, and absorbed knowledge. Distinct from motivated reasoning or confirmation bias in that it is the raw material of the provocation, not the conclusion of the inquiry. The researcher's prior view is the starting blade; the adversarial debate is the whetstone; the refined framework is the sharpened edge.

Adversarial debate

The structured communicative procedure in which a qualified sparring partner, human, AI, or the researcher's own internal critical voice, subjects the provocation or provisional framework to the strongest available objections. The debate is adversarial in the sense that the sparring partner is explicitly tasked with breaking the framework rather than helping it survive. It is productive rather than destructive because both parties share the goal of the best available account of a phenomenon.

Sparring partner

The interlocutor who provides the adversarial challenge in the debate stage. The sparring partner must meet specific qualification criteria: independence from the researcher's prior conclusions, an adversarial disposition (willingness to push the framework to its breaking point), and sufficient domain knowledge or logical rigour to develop objections from the strongest available counterarguments. A sparring partner whose objections are too weak, too sympathetic, or too narrowly drawn produces a debate that functions as validation rather than stress-testing.

Logical debate (Stage 2a)

The first of two sequential debate stages. The logical debate tests the internal coherence of the framework: whether its definitions are consistent, its arguments valid, its implications genuine consequences of its premises, and whether it survives the case that most directly threatens it. The sparring partner at this stage does not require domain expertise. They require adversarial logical rigour and independence from the researcher's conclusions.

Literature debate (Stage 2b)

The second of two sequential debate stages. The literature debate tests the framework's novelty and positioning: whether the construct already exists in the literature, what it adds to adjacent frameworks, and how it should be cited and situated relative to prior work. The sparring partner at this stage requires domain knowledge. The literature functions as a whetstone at this stage — not to construct the framework from the top down, but to sharpen and position a framework that has already survived the logical debate.

The Structural Impossibility Loop

A theoretical construct introduced by Obasa and Obasa (2026) and used as a case study in this paper. It refers to the self-widening gap between what pronatalist policy compensates and what women's expanding aspiration horizons require, driven by the fact that the same prosperity engine that funds policy simultaneously generates new aspirations that make policy insufficient. Used here to illustrate the provocation-to-framework journey that IAI produces.

Confirmed position

A framework that has survived a previous iteration of IAI and represents the current best knowledge on a question. A confirmed position is not a permanent truth. It is the provisional output of a completed cycle, which becomes the starting point for a new cycle when new developments — empirical, theoretical, or technological — challenge it.

Necessity check

Stage Five of the IAI framework. The question: Does this framework add something that did not previously exist? Performed after the logical debate and before or during the literature debate. The necessity check asks not whether the framework is correct but whether it is required — whether existing frameworks can already do what this one does, stated as precisely as it has become through the adversarial process.

3. The Methodological Landscape

IAI does not emerge from nowhere. It is a response to a specific limitation in the existing methodological landscape, and its relationship to that landscape must be stated precisely, both to acknowledge what it owes to existing traditions and to demonstrate where it genuinely departs from them.

3.1 The Hypothetico-Deductive Method and Its Structural Limitation

The hypothetico-deductive (HD) method, foundational to modern social science, proceeds from theory to hypothesis to data collection to confirmation or disconfirmation. Karl Popper (1959) provided its philosophical grounding in the Logic of Scientific Discovery, arguing that the mark of a scientific claim is its falsifiability, i.e., the specification in advance of what evidence would count against it. The HD method has been enormously productive precisely because it imposes discipline: claims that cannot be falsified are excluded from scientific discourse and claims that survive falsification attempts are provisionally accepted.

The method's limitation is structural, not incidental. Popper's falsification criterion requires that a claim be specifiable before it is tested. But some of the most important theoretical contributions in social science involve proposing a claim, or more precisely, a theoretical category, that was not specifiable before the reasoning process that generated it. The Aspiration Horizon is an example: the construct could not have been pre-registered because it did not exist as a named category before the adversarial process that produced it. Asking researchers to pre-register hypotheses before engaging in the kind of reasoning that generates new categories is like asking an architect to submit building permits before knowing what the building is for. The specification presupposes the very thing the inquiry is designed to produce.

Thomas Kuhn (1962) identified this limitation at the level of paradigms, arguing in The Structure of Scientific Revolutions that the most significant advances in science are not produced by normal science, i.e., incremental hypothesis-testing within an existing paradigm, but by paradigm shifts that require a different cognitive activity entirely. Paul Feyerabend (1975) took this further in Against Method, arguing that methodological prescription actively impedes scientific discovery. IAI is not as radical as Feyerabend; it is itself a method, but it takes his observation seriously that the most important new ideas tend to arise through processes that existing methods did not anticipate.

3.2 Grounded Theory

Grounded Theory (GT), developed by Glaser and Strauss (1967) in The Discovery of Grounded Theory, is the qualitative research methodology most often cited as an alternative to hypothesis-first research. GT begins with data collection rather than a prior hypothesis, uses constant comparative analysis to identify codes and categories, and derives theoretical propositions from patterns that emerge in the data rather than imposing them from above.

IAI shares GT's rejection of the hypothesis-first structure and its commitment to generating theory from the inquiry process itself. But the two methods differ structurally in four ways that matter.

First, IAI begins with a provocation; an explicit, post-experiential prior orientation, where GT instructs the researcher to bracket prior assumptions and approach data without preconceptions. IAI argues that bracketing prior assumptions is both impossible and unproductive. The researcher's prior orientation is the raw material of the inquiry, not a contaminant to be suppressed. Making it explicit as a provocation makes it productive rather than merely present.

Second, IAI uses adversarial debate as the primary generative instrument, where GT uses data coding and constant comparison. The framework in IAI emerges from the friction of challenge and response, not from the progressive organisation of empirical data. This means IAI is suited to a different class of problem: generating explanatory frameworks for phenomena where the critical gap is conceptual; the right category does not yet exist, rather than empirical, where the right theory awaits the right data.

Third, IAI has a two-stage debate structure: logical coherence first, followed by literature positioning, which GT does not. GT runs literature review and data analysis iteratively, without a specified sequence. IAI's sequencing is deliberate: a framework that has not survived logical stress-testing should not be subjected to the literature debate, because it will change under that debate, and the positioning will need to be repeated.

Fourth, IAI is explicitly documented. Every challenge raised and every response given is on record. This is partially true of GT through its memo-writing requirement, but GT memos are internal reflections. IAI debate transcripts, particularly when AI is the sparring partner, are records of external challenge and documented response that are reviewable, auditable, and reproducible in ways that internal memos are not.

3.3 The Delphi Method

The Delphi Method, developed at the RAND Corporation by Dalkey and Helmer in the 1950s (Dalkey & Helmer, 1963), is a structured expert consensus procedure in which panellists respond to successive rounds of questionnaires, with anonymised feedback after each round, until consensus converges. It is widely used in health research, policy, and technology forecasting.

IAI shares the Delphi Method's recognition that structured communicative procedure, not solo reflection, is the appropriate instrument for difficult problems under uncertainty. But the two methods differ fundamentally in purpose and direction. The Delphi Method seeks consensus, the convergence of expert opinion toward a shared view. IAI seeks to sharpen the refinement of a provocation through adversarial challenge. The Delphi Method pools expert knowledge to arrive at a collective best estimate. IAI uses adversarial knowledge to test whether a framework can survive the best available objection. These are different epistemological goals and require different procedural designs.

3.4 Abductive Reasoning and Peirce's Self-Correcting Inquiry

IAI's deepest philosophical antecedent is Charles Sanders Peirce's theory of abductive reasoning, developed across his career from the 1870s through the early 1900s (Peirce, 1878, 1903). Abduction, as Peirce defined it, is the logical operation of forming an explanatory hypothesis: observing a surprising fact, and reasoning toward the simplest explanation that would make that fact a matter of course. In his 1903 Harvard Lectures, Peirce stated that abduction "is the only logical operation which introduces any new idea" (Collected Papers 5.172, 1903).

Peirce's mature account of inquiry places abduction first among three stages: abduction generates the hypothesis, deduction derives testable consequences, and induction tests those consequences against evidence. Crucially, Peirce held that this process is self-correcting, that inquiry, continued long enough, progressively approximates truth not by eliminating error but by detecting and revising it through continued engagement with evidence and argument.

IAI makes Peirce's abstract description procedurally concrete. The provocation is the abductive move: the researcher observes a phenomenon that existing explanations do not adequately account for and proposes the simplest explanation that would make it intelligible. The logical debate is the deductive stage: testing whether the hypothesis's consequences hold. The literature debate is the inductive stage: testing the hypothesis against what is known. And the cyclical structure, in which confirmed positions become the starting points of new iterations, is Peirce's self-correcting inquiry given procedural form.

Where IAI extends Peirce is in the specification of the adversarial structure. Peirce described the logic of inquiry but did not specify the communicative procedure for conducting it. IAI does: it names the two debate stages, specifies the qualification criteria for the sparring partner, defines the quality standards that distinguish adequate from inadequate debate, and introduces the documentation requirement that makes the process auditable.

4. The IAI Framework: Full Specification

IAI has five stages. The stages are sequential within a cycle, but the cycle itself is iterative; the output of Stage Five becomes the input of Stage One in the next cycle. Each stage is specified below with its definition, purpose, procedural requirements, and quality criteria.

4.1 Stage One — The Provocation

The provocation is the deliberate, explicit statement of the researcher's post-experimental bias regarding a phenomenon that existing theoretical frameworks inadequately or not at all explain. It is not a hypothesis. A hypothesis is tentative, carefully hedged, and formulated with attention to falsifiability. A provocation is stated with maximum force, overdrawn if necessary, precisely because understatement suppresses the friction that the method requires. Two starting conditions are possible, and the method handles both:

Starting Condition A — The researcher has a bias: The researcher has encountered a phenomenon, formed a view through prior experience and observation, and can explicitly state that view. The provocation is the statement of that view under adversarial pressure. This is the more common starting condition for original theoretical work.

Starting Condition B — The researcher has a confirmed position: A prior cycle of IAI, or some other process of inquiry, has produced a framework that represents the current best knowledge. That framework now serves as the starting point. The provocation for the new cycle is generated by asking: what new development- empirical, theoretical, or technological - challenges this confirmed position? The provocation is then the explicit statement of that challenge against the confirmed position.

The two starting conditions eliminate the apparent chicken-and-egg problem of where inquiry begins. It always begins somewhere, either with a bias derived from experience or with a confirmed position derived from prior inquiry. The method is available at either entry point, and the procedure is the same from either point.

The provocation's quality criterion is precision: it must be stated clearly enough that a qualified sparring partner can identify the strongest objection to it. A vague provocation generates vague objections and vague refinements. The provocation should be uncomfortable for the researcher, the field, or commonly held assumptions, because discomfort indicates that it has identified a genuine gap between what is claimed and what existing frameworks can explain.

4.2 Stage Two — The Adversarial Debate

The debate is the engine of the method. It is divided into two sequential stages, and the sequencing is not incidental; it is a structural requirement.

Stage 2a: The Logical Debate

The logical debate tests the internal coherence of the provisional framework: whether its definitions are consistent, whether its arguments are valid, whether its implications genuinely follow from its premises, and whether it survives the case that most directly threatens it. The sparring partner at this stage does not require domain expertise in the phenomenon under investigation. They require two things: an adversarial disposition, a genuine willingness to push the framework to its breaking point, and independence from the researcher's prior conclusions.

The quality criteria for an adequate logical debate are four:

  • The objections must be sourced from the strongest available counterargument, not from the most convenient one. A debate that only raises weak objections is not a debate; it is confirmation bias with extra steps.
  • The framework must visibly change in response to at least some of the objections. A framework that emerges from the debate, identical to how it entered, has not been stress-tested; it has been validated. Visible change is evidence that the whetstone functioned.
  • The debate must be documented. Every objection raised and every response given must be on record. The documentation is the audit trail that makes the method reproducible and the stress-testing claim verifiable.
  • The debate must be iterative. Each refinement should generate new friction rather than closing the argument. A debate that ends after the first round of objections and responses has not completed the logical stage; it has merely started it.

Stage 2b: The Literature Debate

The literature debate tests the framework's novelty and positioning. It asks: Does this already exist? Where does it sit relative to adjacent frameworks? What does it add that those frameworks do not? The sparring partner at this stage requires domain knowledge, specifically, sufficient familiarity with the relevant literature to identify where the framework duplicates, contradicts, or is superseded by existing work.

The literature functions as a whetstone at this stage, not a blueprint. The framework was not built from the literature; it was built from the provocation and refined through the logical debate. The literature debate uses the existing body of knowledge to sharpen and position the framework that has emerged, revealing where its edges align with and depart from what came before.

The quality criterion for the literature debate is the precision of differentiation: the researcher should be able to state specifically what the framework adds that adjacent frameworks do not, at the level of mechanisms rather than merely outcomes. "This framework is more precise" is not sufficient. "This framework specifies the visibility mechanism through which aspiration horizon expansion suppresses fertility, where adjacent frameworks describe the value shift that results without specifying how it is produced" is sufficient.

4.3 Stage Three — The Stress-Test

The stress test is distinct from the adversarial debate. The debate sharpens the framework through friction. The stress-test asks a more specific question: Is there a case that breaks the framework entirely? Not a case that qualifies it, complicates it, or requires refinement, but a case whose existence would make the framework wrong.

The stress-test is the method's most demanding stage, and it is the one most commonly skipped, not from dishonesty but from the natural human reluctance to seek out the evidence that would destroy one's own creation. The whetstone analogy helps here: a blade that cannot identify the angle at which it breaks cannot be reliably sharpened. The researcher who cannot identify the case that would break the framework cannot be confident that the framework has been genuinely tested.

The stress-test's quality criterion is specificity: the researcher must be able to state, in concrete terms, what evidence or argument would constitute a decisive refutation of the framework. This is not Popper's falsification criterion applied mechanically; IAI does not require that frameworks be falsifiable in the narrow positivist sense. It requires that the researcher has thought hard enough about the framework's limits to know where it stops being applicable, and that they have explicitly stated those limits.

4.4 Stage Four — Implication Pursuit

Once the framework has survived the logical debate and the stress-test, the researcher follows its implications wherever they lead, without filtering for convenience, political comfort, or disciplinary acceptability. The implications are part of the contribution. A framework whose uncomfortable implications are suppressed is a framework whose contribution has been artificially narrowed.

Implication pursuit has one quality criterion: honesty. The researcher must be willing to follow the logic to conclusions that are unwelcome, that contradict prior positions, or that raise questions the framework cannot itself answer. Implications that the framework cannot address should be named as research agenda items or directions for future inquiry, rather than ignored.

4.5 Stage Five — The Necessity Check

Before the framework is formalised for dissemination, the researcher asks: Does this add something that did not previously exist? This is not the same as asking whether the framework is correct. A correct framework that merely restates existing knowledge is not a theoretical contribution. The necessity check asks whether the framework is required, whether the gap it fills was genuinely unfilled before.

The necessity check is performed after the logical debate, because the framework must be precisely enough specified to be compared against the literature in any meaningful way. Comparing a vague intuition against the literature produces a false negative, "this already exists", when in fact what already exists is a related but less precise framework. The precision produced by the adversarial debate is what makes the necessity check actionable.

If the necessity check reveals that the framework already exists, the researcher has two options. Either the framework is genuinely identical to an existing one, in which case the contribution is a new application or extension, not a new framework, or the framework is adjacent but more precise, in which case the contribution is the specification of what existing frameworks leave implicit. Both are legitimate contributions. Neither is a wasted inquiry.

4.6 The Cyclical Structure

IAI is not a linear procedure with a defined endpoint. It is a cycle. The confirmed position that emerges from Stage Five becomes the starting point for a new cycle when new developments, empirical, theoretical, or technological, challenge it. This cyclical structure has three implications.

First, every IAI output is explicitly provisional. The framework that emerges from a completed cycle is the best available account given current knowledge and the quality of the current challenge. It is expected to be superseded when knowledge advances. This is not a concession of weakness. It is an accurate description of how theoretical knowledge actually develops, and it is more honest than methods that present their outputs as final.

Second, the method accumulates. Each iteration produces a framework that is more precisely stress-tested than the one before. The confirmed position that starts Iteration 2 has already survived a debate. The framework that emerges from Iteration 2 has survived two. Over multiple iterations, the researcher builds a progressively refined body of knowledge in which each element carries a documented record of the challenges it has survived.

Third, the method is responsive to advances in knowledge. The trigger for a new iteration is not the researcher's choice but the world's development. When new evidence, a new theory, or a new technology puts the confirmed position under pressure, the cycle begins again. The method is not self-contained. It is in permanent dialogue with what is known.

5. Case Study: The Aspiration Horizon

The development of the Aspiration Horizon framework (Obasa & Obasa, 2026) is presented here not as a retrospective rationalisation but as a worked demonstration of IAI. The framework emerged through the process described in Section 4. The sequence below traces that process stage by stage.

5.1 Stage One — The Provocation

The provocation was stated as follows: “When women want babies, nothing stops them; when they do not, nothing makes them”. Governments, cash transfers, and cultural pressure are irrelevant. The decisive variable is female agency, and right now, in wealthy societies, women do not want babies.

This provocation is deliberately overdrawn. It ignores the evidence that many women in wealthy countries want children but cannot have them for structural reasons. It ignores the distinction between different categories of women: educated versus uneducated, urban versus rural, high-income versus low-income. And it ignores the partial success of Nordic pronatalist policy. These objections were not unknown to the researcher who made the provocation. They were the friction that the provocation was designed to generate.

The post-experimental basis of the provocation was evident: the researcher had absorbed the demographic literature sufficiently to know that something in the standard account was missing, had observed the pattern of pronatalist failure across multiple case studies, and had formed a view about the decisive variable. The provocation made that view explicit and placed it under pressure.

5.2 Stage Two — The Logical Debate

The logical debate produced three refinements that were not in the original provocation.

First, the government-as-irrelevant claim was refined to: governments can only align with the direction women are already moving; they can accelerate or obstruct a trend, but they cannot determine its direction. This is more precise than the original claim and survives the Nordic objection: Nordic policy succeeded because it aligned with the direction educated women were already moving.

Second, the stress test of the refugee camp case: women in objectively dangerous and insecure circumstances have high fertility, which refutes the "fear of the future" explanation that other researchers had offered. The Bunyan Observation: "He that is down needs fear no fall." A woman whose horizon of possible selves is narrow has nothing to lose in parenthood. The child is not in competition with an imagined future because that future has no content. This reframing from fear to aspiration, from safety to horizon width, is the method's most significant output. It arrived not from the literature but from the stress test of an inconvenient case.

Third, the debate produced the two-entry-point structure of the cyclical model: the method works whether the researcher starts with a bias or a confirmed position, and the cycle continues as long as new developments challenge the confirmed position.

5.3 Stages Three Through Five

The stress-test raised five substantive challenges: the Grounded Theory comparison, the replicability objection, the circularity of the necessity check, the bias objection, and the underdetermination challenge. Each was addressed in sequence. The responses produced additional framework specification: the two-stage debate structure, the quality criteria for adequate debate, the whetstone resolution of the bias question, and the cyclical model with two starting conditions. The stress-test record is reproduced in full in Section 8 of this paper.

The necessity checks confirmed that while adjacent frameworks, particularly Lesthaeghe and van de Kaa's (1986) second demographic transition hypothesis, describe a related value shift, none specifies the mechanism by which that shift is produced. The Aspiration Horizon is the first framework to name the visibility mechanism and to identify the role of exposure and imaginability in generating the aspiration that suppresses fertility as the primary operative construct.

6. Cross-Disciplinary Applications

IAI is domain-agnostic. The method does not require a particular kind of data, a particular disciplinary tradition, or a particular type of phenomenon. It requires only that the researcher has encountered a phenomenon that existing theoretical frameworks inadequately explain and has formed a post-experimental view of it. That condition is met across virtually every field of the social sciences and beyond.

6.1 Management and Strategy Research

Most of the important questions in management and strategy are not about data scarcity, as there is abundant data on firm behaviour, market dynamics, and organisational outcomes. The problem is that the categories used to analyse them are increasingly stale. The concept of competitive advantage has not been meaningfully updated since Porter (1985). IAI is well-suited to generating the new category that the existing data has been waiting for, starting from a practitioner's or researcher's post-experiential observation that something in current strategy theory does not fit what they see in the field.

6.2 Public Policy

The gap between what the evidence shows and what the policy debate addresses is often not empirical but categorical: policymakers are asking the wrong question because they do not yet have the concept to ask the right one. IAI generates concepts. The provocation-to-framework journey it traces is precisely the journey from "something is wrong with how we are thinking about this problem" to "here is the category that makes the problem legible."

6.3 Clinical and Social Psychology

The replication crisis in psychology is partly attributable to researchers’ testing hypotheses derived from unexamined categories. The method's distinctive contribution in this context is the logical debate stage: requiring that the internal coherence of a theoretical construct be subjected to adversarial pressure before the construct is operationalised and tested empirically would significantly reduce the proportion of studies that produce statistically significant results for constructs that are theoretically incoherent.

6.4 Economics

The most intellectually productive recent work in economics, behavioural economics, complexity economics, and feminist economics has come from researchers who noticed that the existing categorical apparatus did not fit what they observed and insisted on building new categories rather than forcing the data into old ones. IAI formalises that instinct, making it available to researchers who have the observation but not the confidence to trust their own departure from the field's dominant categories.

6.5 Artificial Intelligence Research

The speed at which new phenomena are appearing in AI research has outpaced the development of adequate categories to describe them. Terms like "alignment," "hallucination," and "emergence" are in widespread use without rigorous theoretical definition, because the field has not yet produced frameworks adequate to the phenomena it has created. IAI, particularly with AI as the sparring partner, allowing the method to run at the speed of the phenomena themselves, is unusually well-suited to generating those frameworks.

7. Artificial Intelligence as Sparring Partner

The emergence of large language model artificial intelligence (AI) as a thinking interlocutor is the most significant development in the practical availability of IAI since the method was first practised, informally and without a name, by researchers working through arguments with well-read colleagues. It warrants a dedicated section.

7.1 What AI Changes

The adversarial debate stage of IAI requires a sparring partner who meets specific qualification criteria: independence from the researcher's prior conclusions, an adversarial disposition, and sufficient knowledge to raise objections grounded in the strongest available counterarguments. For most researchers working alone, without a well-informed colleague available in real time, willing to push hard against their ideas, this requirement was effectively a barrier. The method existed as an ideal that could only be approximated by the most well-connected researchers in the most intellectually active institutional environments.

AI changes this. A well-trained large language model can source objections from across disciplines simultaneously, maintain an adversarial disposition without the social friction that makes human colleagues reluctant to push too hard, generate challenges in real time rather than over weeks of correspondence, and document every exchange by default. The IAI documentation requirement, previously aspirational for most solo researchers, is automatically met when AI is the sparring partner. The transcript exists. The audit trail is complete. The exchange is reproducible.

7.2 What AI Does Not Change

AI changes the accessibility and documentation of the debate stage. It does not change the method's other stages, nor does it change the most important variable in the method's output: the quality of the provocation and the intellectual courage of the researcher who creates it.

AI can source the strongest available objections from the literature it has been trained on. It cannot source objections that are not in its training data. The unknown-unknown problem, the objection that neither the researcher nor the AI has encountered, remains. The method's response to this limitation is the one we have already established: the absence of an objection does not prove the framework is unbreakable. It proves it has survived the best available challenge. This is what the Aspiration Horizon case study demonstrates: the framework produced is the best available account given the knowledge and challenge available at the time of its production. It is expected to be refined when new challenges emerge.

The quality of the provocation remains entirely human. AI can sharpen a blade. It cannot choose what to make the blade from. The post-experiential bias that initiates the provocation, the researcher's prior encounter with the phenomenon, the observation that something in existing frameworks does not fit, is the irreducibly human entry point. No method generates this. Every method assumes it.

7.3 Author Note on AI Use in This Paper

The Aspiration Horizon framework and the IAI methodology were developed through structured adversarial dialogue with Claude (Anthropic), an AI assistant. The theoretical frameworks, their defences under stress-testing, the naming, and all substantive intellectual positions are the authors'. The AI contributed the stress-test challenges, philosophical literature connections, prose drafting, and document formatting. The authors take full responsibility for all claims. A full transcript of the adversarial dialogue from which both frameworks emerged is available from the corresponding author on request. This paper is itself a demonstration of the method it describes.

8. The Stress-Test Record

Transparency requires that the challenges raised against the IAI framework during its development be presented in full, alongside the responses that survived. This section reproduces the five substantive challenges and their resolutions. The record is presented here not as an appendix but as a core section of the paper, because the method's claim, that adversarial debate produces stronger frameworks, can only be evaluated by seeing the debate.

8.1 Challenge One: IAI Is Just Grounded Theory With Extra Steps

The objection: Glaser and Strauss (1967) already said: do not start with a hypothesis, let theory emerge from the inquiry process, use constant comparison to refine categories. IAI appears to replicate this structure under a different name.

The response: IAI and Grounded Theory share the rejection of hypothesis-first structure but differ structurally in four ways that matter for output quality. IAI begins with an explicit post-experiential provocation rather than bracketing prior assumptions. IAI uses adversarial debate rather than data coding as the primary generative instrument. IAI has a specified two-stage debate sequence that GT does not. And IAI produces debate transcripts as its primary documentation, which serve as auditable records of external challenges rather than internal memos. These differences are not stylistic. They produce different outputs for different problem classes. GT is designed for phenomena where the gap is empirical, where the right theory awaits the right data. IAI is designed for phenomena where the gap is conceptual, where the right category does not yet exist. The methods are complementary, not identical.

8.2 Challenge Two: The Friction Is Not Replicable

The objection: the method's quality depends entirely on the sparring partner's quality. A weak sparring partner produces a weak framework. Without specifying what qualifies a sparring partner, the method is not a procedure; it is advice.

The response: the sparring partner qualification criteria are specifiable and have been specified. The sparring partner must be independent of the researcher's prior conclusions, have an adversarial disposition, and have sufficient knowledge to raise objections grounded in the strongest available counterarguments. The debate's adequacy is assessable through four quality criteria: the strength of the objections sourced, visible change in the framework, documentation of the exchange, and an iterative structure. The PhD defence analogy is instructive: the jury's composition is governed by rules: expertise requirements, independence requirements, and disciplinary range requirements. IAI's sparring partner specification is the equivalent of those jury composition rules.

8.3 Challenge Three: The Necessity Check Is Circular

The objection: confirming that a framework is novel requires comprehensive literature knowledge, but the method explicitly does not require comprehensive prior literature knowledge. You cannot confirm novelty without the knowledge that would have prevented the framework from being generated. The response: the objection assumes that provocation and literary knowledge are mutually exclusive. They are not. The provocation is post-experiential. It is generated from prior absorbed knowledge, not from a vacuum. Bias cannot exist without a prior knowledge base. The literature does not construct the framework; it serves as a whetstone to sharpen and position a framework that has already been generated through provocation and refined through logical debate. The necessity check, performed after the logical debate, uses a targeted comparative search, comparing a precisely specified, stress-tested framework with the literature rather than comprehensive prior knowledge. Comparing a precise framework against the literature asks a much more answerable question than comparing a vague intuition.

8.4 Challenge Four: The Method Confirms What the Researcher Already Believed

The objection: the provocation is not neutral. It is derived from the researcher's prior worldview. The method provides no mechanism to detect when the initial bias has survived the debate without being genuinely tested.

The response: the debate quality criteria already contain the answer. A debate in which the framework does not change is not a debate; it is confirmation bias with documentation. The quality criterion requiring a visible change in the framework is the bias-detection mechanism. Furthermore, the documented output carries its own evidence: a framework that has been genuinely stress-tested looks different from one that has been validated. It carries acknowledged limitations, explicit scope boundaries, refined definitions, and a record of the points at which the researcher's original position was revised. The Aspiration Horizon carries all of these marks: the acknowledgement that individual psychological process is beyond its scope, the two-stage debate structure that emerged from stress-testing, and the refined distinction between transactional and foreclosed-self costs. The marks of friction are evidence of genuine sharpening.

8.5 Challenge Five: The Method Is Underdetermined

The objection: the five stages specify dispositions rather than procedures. Without a granular procedural specification at every decision point, the method is not replicable; it describes how good theorists think, not a procedure that makes thinking better.

The response: The level of specification required of a methodology depends on what it is designed to produce. A Likert scale specifies response format precisely because the measurement instrument must be identical across administrations for the results to be comparable. IAI specifies debate quality criteria, sparring partner requirements, documentation standards, and a two-stage structure — sufficient specification to make the method teachable and reproducible. What it does not specify is Stage One: how the researcher identifies the observation worth provoking. But no method specifies this. The Likert scale does not tell the researcher what attitude to measure. Grounded Theory does not tell the researcher which phenomenon to study. Every method assumes a prior act of recognition: someone noticing something worth investigating. IAI makes this assumption explicit rather than concealing it, which is consistent with the method's general treatment of prior orientation as productive rather than suppressed.

9. Conclusion

This paper has introduced Iterative Adversarial Inquiry; “the Whetstone Method”, as a framework for generating original theoretical contributions through structured adversarial debate. The central argument is that hypothesis-first research, for all its virtues, has a structural blind spot: it cannot generate the theoretical category that does not yet exist, because you cannot pre-register a hypothesis about something you do not have a name for. IAI is designed for precisely that class of problem.

The method's five stages: provocation, adversarial debate in two sequential stages, stress-test, implication pursuit, and necessity check, together produce a framework that has been sharpened rather than insulated, challenged rather than protected, and documented rather than asserted. The cyclical structure ensures that the method remains permanently responsive to advances in knowledge: confirmed positions become the starting points of new cycles when new developments put them under pressure.

The Aspiration Horizon case study demonstrates that the method works. It produced a framework that is more precise than the intuition that initiated it, more defensible than one constructed through hypothesis-first research, and more honest about its own limitations than one that was never genuinely challenged. The stress-test record in Section 8 makes this claim verifiable: the challenges and the responses are on record.

The emergence of AI as a sparring partner extends the method's availability from the well-connected researcher with an ideal interlocutor to any researcher with a genuine observation and the intellectual courage to have it challenged. That extension is not a dilution of the method. The quality criteria remain. The sparring partner requirements remain. The documentation standard is, if anything, higher when AI is the interlocutor, because the transcript exists by default. What changes is access. IAI was always available to the Keynes who had the Cambridge common room. It is now available to the researcher who works alone.

The method does not give a researcher the instinct to notice something worth investigating. No method does. What it gives is a procedure for turning that instinct into a contribution that can be evaluated, challenged, built upon, and refined by others. That is what methodology is for. That is what the Whetstone Method does.

10. References

All references are formatted in American Psychological Association (APA) 7th edition style.

Bellucci, F., & Pietarinen, A.-V. (2023). Peirce's abduction. In L. Magnani (Ed.), Handbook of abductive cognition (pp. 147–172). Springer. https://doi.org/10.1007/978-3-031-10135-9_7

Charmaz, K. (2006). Constructing grounded theory: A practical guide through qualitative analysis. SAGE Publications.

Corbin, J., & Strauss, A. (2008). Basics of qualitative research: Techniques and procedures for developing grounded theory (3rd ed.). SAGE Publications.

Dalkey, N., & Helmer, O. (1963). An experimental application of the Delphi method to the use of experts. Management Science, 9(3), 458–467. https://doi.org/10.1287/mnsc.9.3.458

Douven, I. (2017). Abduction. In E. N. Zalta (Ed.), Stanford Encyclopedia of Philosophy (Summer 2017 ed.). Stanford University. https://plato.stanford.edu/archives/sum2017/entries/abduction/

Feyerabend, P. (1975). Against method: Outline of an anarchistic theory of knowledge. New Left Books.

Frankfurt, H. G. (1958). Peirce's notion of abduction. The Journal of Philosophy, 55(14), 593–597. https://doi.org/10.2307/2022448

Glaser, B. G., & Strauss, A. L. (1967). The discovery of grounded theory: Strategies for qualitative research. Aldine Publishing Company.

Glaser, B. G., & Strauss, A. L. (1965). Awareness of dying. Aldine Publishing Company.

Kuhn, T. S. (1962). The structure of scientific revolutions. University of Chicago Press.

Lakatos, I. (1978). The methodology of scientific research programmes. Cambridge University Press.

Lesthaeghe, R., & van de Kaa, D. J. (1986). Twee demografische transities? [Two demographic transitions?]. In D. J. van de Kaa & R. Lesthaeghe (Eds.), Bevolking: Groei en krimp (pp. 9–24). Van Loghum Slaterus.

Likert, R. (1932). A technique for the measurement of attitudes. Archives of Psychology, 22(140), 1–55.

Linstone, H. A., & Turoff, M. (Eds.). (1975). The Delphi method: Techniques and applications. Addison-Wesley.

Obasa, O., & Obasa, P. (2026). The aspiration horizon: Why wealthy societies cannot buy their way back to replacement fertility. SSRN Working Paper Series. [DOI pending]

Peirce, C. S. (1878). Deduction, induction, and hypothesis. Popular Science Monthly, 13, 470–482.

Peirce, C. S. (1903). Pragmatism and abduction. Lecture VII of the Harvard Lectures on Pragmatism, May 14, 1903. In C. Hartshorne & P. Weiss (Eds.), Collected papers of Charles Sanders Peirce (Vol. 5, §171–172). Harvard University Press (1934).

Peirce, C. S. (1933–1958). Collected papers of Charles Sanders Peirce (Vols. 1–8; C. Hartshorne, P. Weiss, & A. Burks, Eds.). Harvard University Press.

Popper, K. R. (1959). The logic of scientific discovery. Hutchinson. (Original work published 1934 as Logik der Forschung)

Porter, M. E. (1985). Competitive advantage: Creating and sustaining superior performance. Free Press.

Strauss, A., & Corbin, J. (1990). Basics of qualitative research: Grounded theory procedures and techniques. SAGE Publications.

11. Author Note

Olorunfemi Obasa, MBA and Precious Obasa (née Olusegun), PhD, are Cofounders at Praxis by Extrafemi, a project of Extrafemi, Nigeria. Both authors contributed to the conceptualisation of the Iterative Adversarial Inquiry framework and are jointly responsible for all claims in this paper. This paper was developed through structured adversarial dialogue with Claude (Anthropic), an AI assistant that served simultaneously as the framework's stress-testing sparring partner and drafting assistant. The theoretical framework, its conceptual architecture, the naming decisions, and all substantive intellectual positions defended throughout the stress-test are the authors'. The AI contributed to identifying adversarial challenges, identifying philosophical literature, drafting prose, and formatting documents. The authors take full responsibility for all claims made in this paper. The production of this paper itself demonstrates the Iterative Adversarial Inquiry method it describes. A full transcript of the adversarial dialogue from which both the Aspiration Horizon framework and the IAI methodology emerged is available from the corresponding author on request. The transcript constitutes the primary audit trail for the contribution percentage assessment below.

Estimated contributions:

Olorunfemi Obasa, MBA: 55% (conceptualisation, framework development, stress-test responses, naming, intellectual positions).

Precious Obasa (née Olusegun), PhD: 15% (conceptualisation, validation).

AI (Claude, Anthropic) — 20% (adversarial challenges, literature identification, prose drafting, document formatting).

The authors have no conflicts of interest to declare.

— End of Working Paper —

Cite this

Obasa, O., & Obasa, P. (2026). Iterative Adversarial Inquiry: The Whetstone Method (Extrafemi Research Paper No. 26-002). SSRN. https://doi.org/10.2139/ssrn.6659879