Concepts, open questions and sources from Nathan Sweet’s work.
Worldview
Every person, instrument and model encounters the world through a limited perspective. My working worldview asks how those perspectives can correct one another. It gives strong local findings their due, keeps wider interpretations answerable to evidence, and leaves room for new categories when the world exceeds the map.
Our senses, language, instruments and models reveal some things and miss others. Acknowledging that limitation does not make every opinion equal. It asks us to check where a conclusion travels reliably, seek perspectives that expose its blind spots, and keep the strength of the claim matched to its evidence.
Question: How can a limited view still yield reliable knowledge?
Reconsider if: Revise a stated limit or scope restriction when a reliable method demonstrates access or transport the current account denied.
A surviving text or familiar category reaches us through selection, recording, translation and institutions. Reconstructing that route can reveal a hidden assumption. An older or suppressed source is not automatically truer; the history matters when it changes what the evidence supports.
Question: What has the route to us changed?
Reconsider if: Simplify the procedure when additional genealogy fails to alter warranted interpretation, prediction or action under a fair cost comparison.
A tested physical mechanism earns credit for what it explains. It need not settle what all reality is made of. I use naturalistic rivals to ask whether a further source contributes anything distinctive, while applying the same demand to my own stronger claims.
Question: What does a naturalistic explanation actually earn?
Reconsider if: Withdraw a claim of explanatory adequacy if the reconstruction loses target-relevant features or a rival earns a defended comparative advantage.
A possibility can remain unresolved while I decline to make it part of my working worldview. Not proving a claim false does not make it an earned explanation. Equally, rejecting its current warrant must leave open the evidence or argument that could justify accepting it.
Question: Must an unrefuted possibility become a belief?
Reconsider if: Reconsider rejection when the proposal earns a claim-appropriate advantage over serious alternatives, or when the criticism misstates its content.
A person, model, memory system and tools can solve a problem together that none solves alone. The capability should first be attributed to the combination that produced it. That does not establish a single conscious mind or a single responsible agent spanning the whole arrangement.
Question: Where is the capability actually produced?
Reconsider if: Weaken system-level attribution when removing the alleged coupling leaves the relevant capability intact or a simpler component model explains the gain.
A person need not contain an unchanging inner object to have real continuity. Bodies, habits, memories, relationships and recurring organization can carry a history forward. The question is what kind of continuity matters for this person, system or decision, rather than demanding one rule for everything.
Question: What persists when the parts change?
Reconsider if: Revise the proposed criterion when branching, replacement or disruption cases show that it misclassifies the continuity relevant to the target.
Methodology
The Recursive Corrigibility Framework makes the weak joints of an argument easier to inspect. State what is being claimed, compare serious alternatives, identify what would count against each, then correct the particular assumptions that failed. The method is a developing proposal, and its extra machinery must earn its cost.
A mathematical proof, a historical claim and a moral judgment cannot all be checked in the same way. First identify what the sentence asks us to accept. Then use the evidence and argument suited to it, keeping assumptions, measurements and interpretations separate.
Question: What would count as a good reason here?
Reconsider if: Split, merge or replace a type when it obscures the relevant target or repeatedly routes inquiry to an inappropriate standard.
Before declaring an explanation unnecessary, check that its replacement still explains the thing under dispute. Removing a word can also remove a relationship or a lived concern. If the pieces cannot be separated fairly, compare the strongest reconstructed accounts as wholes.
Question: Did the alternative preserve the question?
Reconsider if: Abandon the subtraction and compare reconstructed wholes when removing the disputed commitment changes the target or its constitutive relations.
Two stories can both fit a result. A useful next test is one that would change their relative support, rather than merely produce another example both can explain. Name the outcomes, relevant controls and possible failures before seeing which result appears.
Question: What should come out differently?
Reconsider if: Revise the test when its outcomes do not separate the intended rivals or when its implementation violates the assumptions that made it discriminating.
A failed explanation need not erase the observation, and one incorrect sentence need not destroy an entire framework. Trace which assumptions produced the mistake, update the claims that depend on them, and preserve independent reasons that still hold.
Question: How far should a failure travel?
Reconsider if: Broaden or narrow the correction when an audited support path shows that the original propagation missed a real dependency or removed an independent warrant.
Saving a correction is useful only if it can influence the next relevant decision. Keep the failed claim, why it failed and what should reopen it. A memory system should also let new evidence correct the correction, rather than turn yesterday’s lesson into a permanent rule.
Question: Can yesterday’s mistake change tomorrow’s action?
Reconsider if: Reject a claimed learning gain when matched future tasks show no reliable improvement, harmful overgeneralization or no advantage over simpler memory.
RCF is a proposal for better reasoning, not an exemption from criticism. More terminology and more checkpoints can make work slower without making it better. Compare the full procedure with simpler alternatives, including what people can understand, correct and use afterward.
Question: Does this machinery improve inquiry?
Reconsider if: Compress or remove operators when matched held-out comparisons show no material gain or a worse error, accessibility or cost profile.
Ontology
A cell, a person and an institution can each be real at the scale where their organization makes a difference. Boundary- and Scale-Relative Realism asks which patterns earn prediction, causal identification and transfer, rather than declaring one size of thing uniquely real. Its boundaries and categories are hypotheses too.
A pattern deserves serious commitment when it repeatedly helps us predict, intervene or transfer understanding to new cases better than serious alternatives. Usefulness is evidence to examine, not an automatic certificate that the pattern reveals reality’s intrinsic nature.
Question: When is a pattern more than a convenient story?
Reconsider if: Weaken realism about a particular pattern when its advantage disappears under transfer or better identified rivals explain the result without that pattern.
Cells, organisms and institutions can make different questions tractable. A higher-level explanation can be valuable even when its components are physically realized. The scale must earn its usefulness for the target; neither being smaller nor being larger grants automatic explanatory authority.
Question: Why should the smallest scale get every explanation?
Reconsider if: Prefer a different scale when it preserves the target and consistently improves explanation, identified intervention or transfer under a fair comparison.
We often draw a line around an organism, a device or a person before asking what produces its capability. Some problems require a wider or narrower boundary. Test which components and relationships matter through removal, replacement and recovery, rather than letting a familiar outline decide.
Question: Which boundary does the work?
Reconsider if: Redraw or reject the candidate boundary when perturbation and substitution show the attributed organization lies elsewhere or the added scope earns no gain.
Energy, materials and implementation matter whenever something is physically done. Those constraints do not by themselves explain the full organization or settle what consciousness is. Keep the particular operation and measurement visible, rather than asking a broad principle to answer questions it never tested.
Question: What does physical realization leave open?
Reconsider if: Revise an application when its operation, system boundary, units or physical assumptions fall outside the principle’s justified domain.
Reasoning, agency, self-report, feeling and welfare are related questions, not interchangeable achievements. Investigate each without declaring one familiar marker the definition of consciousness. Reports matter, but what produces them and how they respond to context also matter.
Question: Does successful control establish experience?
Reconsider if: Revise a phenomenal attribution or marker when matched perturbations dissociate it from the relevant capacities, or a better theory explains the evidence with fewer unsupported bridges.
A system can respond to information, revise a policy and alter what happens next while inheriting the conditions that made it capable of choosing. Ask which reachable outcomes its organization helps control. Keep that practical control separate from ultimate authorship, consciousness and moral responsibility.
Question: Can control be real without ultimate self-creation?
Reconsider if: Weaken the agency attribution when interventions show no target-relevant leverage, the proposed feasible alternatives are not reachable, or a simpler passive model accounts for the result.
Ethos
Sweet Rationalism asks for generous understanding and demanding criticism, including criticism of the framework itself. It keeps values visible, corrections usable and affected people able to contest how their interests are represented. The point is to leave a disagreement with better understanding, a better question or a better decision.
Before criticizing a view, preserve its question, qualifications and strongest supported meaning. Quoting a memorable fragment can be misleading when the surrounding words change its force. Fair reconstruction makes criticism more demanding because it has to address the real disagreement.
Question: Would the other person recognize the argument?
Reconsider if: Revise the reconstruction when verified context or a justified source correction shows that it alters the original claim’s content or scope.
Take a rival seriously without pretending all claims have the same evidence. Different histories and mechanisms can justify different expectations. Fairness asks for an identifiable reason when relevantly matched evidence receives different treatment, including when the favored conclusion is mine.
Question: Must every possibility start equally likely?
Reconsider if: Change the comparison when an identified evidence or mechanism difference defeats the match, or when the symmetry procedure systematically worsens calibration.
Experiments can estimate what happens. Proofs establish what follows under premises. Historical records, experience and ethical reasoning contribute different kinds of warrant. Inquiry becomes unfair when it demands the wrong kind, or allows a successful answer to one question to settle another.
Question: Which kind of answer does the question need?
Reconsider if: Reroute the claim when the chosen standard excludes relevant evidence, grants immunity or fails to test the commitment actually asserted.
When someone can meaningfully explain what they value or find harmful, that evidence deserves a hearing. It is neither infallible nor disposable. Decisions also need reasons about consent, power, conflicting interests and who has authority to act.
Question: Who gets to define another’s good?
Reconsider if: Revise the weight or procedure when calibrated evidence reveals incapacity, manipulation, serious instability or harms the current arrangement fails to address.
A finding can tell us what a decision may do without deciding which costs are acceptable. State the values, who benefits, who bears the risk and what alternatives are available. Uncertainty can justify proportionate care while the deeper metaphysical questions remain open.
Question: What justifies the action we recommend?
Reconsider if: Revise the action when its normative bridge cannot be defended, relevant affected interests were omitted, or a feasible alternative better respects the justified commitments.
A correction should tell readers what changed, what remains and why. It should be possible to disagree without losing the right to participate. The same standard applies to Sweet Rationalism: if its language protects the project or makes inquiry unusable, the practice needs to change.
Question: Does the correction reach the next decision?
Reconsider if: Simplify or replace the practice when it adds social or cognitive burdens without reducing material misunderstanding or improving later decisions.
Framework Bridge
An analogy, a translation or a shared equation can be useful while hiding a changed meaning. The Framework Bridge asks what a claim means in both places, what the mapping preserves, what it loses and what would expose a failed crossing. Agreement should survive a return to the original sources.
Two theories can use the same word for different things, or different words for a similar role. A useful translation records what each term refers to, which inferences it licenses and what changes under the mapping. Similar vocabulary alone does not establish agreement.
Question: Does the same term do the same work?
Reconsider if: Revise or suspend the translation when it changes a load-bearing inference or informed source comparison reveals a material mismatch.
An account can look simpler because its hardest question disappeared in translation. Before subtracting a disputed commitment, check that the replacement preserves the relevant phenomenon, relationship and practical concern. If it does not, compare the accounts without pretending their common core was secured.
Question: Did the translation erase what mattered?
Reconsider if: Reject the claimed shared core when reconstruction loses target-relevant organization or changes the question it purports to answer.
A useful summary always leaves something out. The important distinction is whether the missing detail changes the question, prediction or decision. Keep those losses visible so a smoother explanation does not create agreement that vanishes when readers return to the original accounts.
Question: What disappeared on the way across?
Reconsider if: Restore or expose a distinction when its omission changes a relevant inference, outcome, value judgment or interpretation of evidence.
Different sensors, categories and interfaces can transform the same target in different ways. Before reading the disagreement as a different reality, examine what each route measured and omitted. A complete reconstruction may be impossible, but enough transparency can reveal which dispute is actually live.
Question: Are we disagreeing about the target or the instrument?
Reconsider if: Revise the proposed reconstruction when calibration, held-out transformations or negative controls show that it cannot recover the claimed invariant.
A reliable experiment at the start of an argument cannot guarantee every conclusion several steps later. At each crossing ask what new assumption entered, what still needs support and which rivals retain the result. Keep a real discovery even when its wider interpretation overreaches.
Question: How far can one strong result travel?
Reconsider if: Narrow a remote conclusion when a load-bearing bridge lacks warrant, while checking for independent support before withdrawing it entirely.
The atlas should store claims, their sources, dependencies, serious rivals and what could change them. An arrow is a proposed relationship to inspect, not proof that the relationship is true. When something fails, the map should help locate which other claims actually depend on it.
Question: What does an arrow in the graph establish?
Reconsider if: Change or remove an edge when its stated relation is unsupported; simplify the map when it adds structure without improving understanding or correction.
Mind & consciousness
Intelligence, agency, consciousness, feeling and selfhood are connected questions, but they are not interchangeable. These essays ask which capacities the evidence supports, which subject or system they belong to, and what remains open when a theory explains only some of them.
A system might learn, pursue a goal, describe itself or coordinate a group. Those observations can support particular capacities without settling whether it feels anything, whether it can suffer or where a subject begins. Naming the target first lets evidence bear on the question it can actually answer.
Question: Does success on one capacity settle the others?
Reconsider if: Merge or relate the targets more strongly if a defended identity argument or discriminating evidence establishes the connection.
The hard-problem essays ask what remains after questions about experience, reports, organization, subject boundaries and explanatory bridges are distinguished. Their proposal is conditional: a unitary problem dissolves only if the separation leaves no further positive target and loses no necessary explanatory relation.
Question: After decomposition, is there an additional positive target?
Reconsider if: Withdraw the dissolution claim if a coherent positive residual or necessary lost relation survives the strongest reconstruction.
First-person reports are evidence produced in a situation. Attention, memory, vocabulary, incentives and reporting policies can shape them. The same caution applies to people and computational systems, while their different evidence and histories still matter. Reports should be examined rather than granted certainty or discarded by category.
Question: How much can a self-report tell us?
Reconsider if: Change the evidential weight when calibration, perturbation or provenance reveals a different relationship between report and target.
A person and their tools may form a useful unit for explaining a capability. That does not establish one experiencing subject spread across all the parts. The boundary for control, the boundary for responsibility and the boundary for experience can differ.
Question: What follows from distributed cognition?
Reconsider if: Revise the boundary when intervention, dissociation or a defended theory shows which organization is necessary for the target capacity.
Agency & ethics
A system can affect others, make choices, or depend on our actions in very different ways. Ethical judgment needs an account of whose interests are affected, what grounds protection, who has authority to act and how uncertainty should change the decision.
Damaging a component can harm an already protected person, animal or ecological system. That gives a reason to constrain the action even if the component has no interests of its own. The conclusion concerns the protected bearer and the damaging action; direct standing for the component needs a further argument.
Question: Can something matter without feeling?
Reconsider if: Withdraw the particular constraint if the relevant constitutive, impairment or normative premise fails, or justified competing duties defeat it.
Direct moral standing asks whether a system matters for its own sake. A defensible answer needs a reason why a feature warrants protection and evidence that this particular system has that feature. Agency, usefulness, a familiar body or a compelling description cannot silently supply both parts.
Question: What would justify direct moral standing?
Reconsider if: Revise the attribution when its moral ground cannot be defended or its claimed instantiation fails target-appropriate assessment.
A system can exercise local control without being the ultimate author of its own history. Responsibility also depends on role, foreseeability, prevention and authority. Ability alone gives no permission to intervene, and a lack of ultimate self-creation need not erase practical accountability.
Question: Who can act, who may act, and who answers?
Reconsider if: Change the attribution if a better account of control, authority or obligation defeats the premises used in the particular case.
An unresolved theory of mind does not remove every reason for care. Decisions can consider credible harms, reversibility and affected interests before ontology is settled. Precautions also have costs, so they need comparison with feasible alternatives and the risks of exclusion.
Question: How should uncertainty change what we do?
Reconsider if: Revise the safeguard as evidence, severity, reversibility or feasible alternatives change the justified balance.
Metaphysics & religion
An experience, an organism or a mathematical pattern can be real while its wider interpretation remains disputed. These articles compare the work done by proposed sources, including God, universal experience and Platonic ingress, with what serious alternatives can explain.
When rival accounts both accommodate a result, the result does not by itself select either account’s distinctive addition. Ask what the further posit explains, predicts or otherwise warrants beyond the strongest rival. The relevant gain may be conceptual or explanatory as well as experimental, depending on the claim.
Question: Which part of the evidence favors the extra source?
Reconsider if: Reassess the comparison when a target-preserving rival gains a warranted advantage, or a defended independent argument establishes additional content.
The Whitehead essays ask which explanatory roles require subjective prehension or divine valuation, and which can be recovered through organization, inherited constraints and causal processes. A successful reconstruction challenges a necessity claim only where it preserves the original target.
Question: Does the explanation require its metaphysical interpretation?
Reconsider if: Withdraw the necessity criticism where the reconstruction loses a warranted target or an independent argument establishes the contested commitment’s role.
A profound experience deserves serious description. Its vividness does not alone identify a deity, a nonphysical realm or the intrinsic nature of everything. The interpretation needs reasons that distinguish it from alternatives while preserving what the participant actually encountered.
Question: What does an encounter establish about its cause?
Reconsider if: Revise the interpretation or its criticism when discriminating evidence or a stronger reconstruction changes what the experience supports.
A reply can be long and sophisticated while addressing a position the critic did not hold. Put the exact claims side by side, preserve qualifications, and ask which inference the reply defeats. The same rule applies when the alleged misrepresentation turns out to be the critic’s own mistake.
Question: Does the reply reach the original objection?
Reconsider if: Withdraw a misrepresentation charge if the attributed position is supported in context or the reply validly establishes the claimed implication.
Biology & computation
Bioelectricity, xenobots and self-sorting arrays raise worthwhile scientific questions. The task is to preserve what the experiments show while checking which mechanisms, mathematical claims and stronger interpretations their evidence actually supports. Public code makes some parts unusually open to inspection.
Bioelectric interventions and unusual cellular collectives can be scientifically valuable whether or not a Platonic interpretation succeeds. Keep the measurements, mechanisms and limits visible. Additional claims about nonphysical causes, minds or welfare need their own supporting bridge.
Question: What remains if the metaphysical interpretation is removed?
Reconsider if: Revise the credit assignment when a discriminating test or defensible independent argument identifies the additional contribution.
The sorting debate contains separate questions about how clustering forms, how tie handling may retain it, and what a random benchmark predicts. An endpoint expectation does not determine the entire trajectory. A proposed mechanism must match the actual sorting policies rather than textbook assumptions alone.
Question: What generates the clusters, and what keeps them?
Reconsider if: Change the proposed local mechanism when controlled, implementation-matched comparisons show it fails to produce or retain the relevant effect.
A sequential simulator, an operating-system-threaded implementation and a biological experiment are different systems. Reproducing one is useful, but the step to the others needs a checked mapping of rules, scheduling, measurements and conditions. A repository date alone does not identify the code behind every figure.
Question: Which system did we actually test?
Reconsider if: Narrow or withdraw a generalization when a load-bearing implementation difference or missing provenance defeats the claimed correspondence.
Public scripts, proofs and citation checks make claims easier to inspect. Their outputs still depend on assumptions, source identity and acceptance rules. Multiple computational reports can share an error. A literal verification label is not a substitute for a check that can fail.
Question: What did the validator actually validate?
Reconsider if: Correct or withdraw an audit finding when a reproducible countercheck defeats it, preserving independent findings that remain supported.
- My current worldview
Every person, instrument and model encounters the world through a limited perspective. My working worldview asks how those perspectives can correct one another. It gives strong local findings their due, keeps wider interpretations answerable to evidence, and leaves room for new categories when the world exceeds the map.
- When criticism is dismissed without examples
Examines Michael Levin’s public comments about criticism assisted by language models. Sweet agrees that volume and repeated arguments can make engagement difficult, then asks what specific errors or unanswered sources support dismissing a particular criticism. The article distinguishes the origin of an argument from its content and calls for checkable examples.
- Xenobots: the biology and the extra claim
Introduces xenobots and distinguishes the different experiments grouped under that name. It takes their unusual behavior seriously while asking what inherited cellular machinery, assembly, environment and collective dynamics explain. Sweet argues that surprise or a missing task instruction does not by itself identify an additional nonphysical source of organization.
- What does a God interpretation actually add?
Responds to Matthew Segall’s account of Whitehead by following the explanatory work assigned to God, universal experience and divine valuation. Sweet preserves common ground about process and relationship while asking which additional commitments improve the explanation. The discussion also keeps the sufferer, rather than only a cosmic interpretation, in view.
- Divine love, suffering and the prey
Tests a process theology of divine love against suffering, predation and the interests of those harmed. Sweet asks what remembering or incorporating suffering accomplishes for its original bearer, and whether a more limited God improves the comparison. Religious experience and consolation remain distinct from identifying their proposed cosmic source.
- Naturalizing Whitehead’s God
Develops a naturalistic comparison with the causal and practical roles attributed to Whitehead’s God. The article asks which roles can be reconstructed through organization, history, selection and collective inquiry, and which remain unresolved. Sweet applies the same scrutiny to his own account and proposes conditions under which the comparison should change.
- What follows from no direct consciousness measure?
Examines the claim that science lacks evidence for consciousness and asks what follows from that limitation. Sweet distinguishes direct access, indirect indicators, metaphysical interpretation and comparative testing. The article proposes precise questions and risky comparisons while warning that uncertainty cannot automatically award support to a preferred account of experience.
- Bubble sort and the missing causal inference
Uses self-sorting arrays to separate emergent behavior from claims about its ultimate source. Sweet examines the published rules, interviews and Joel Dietz’s analysis, including chance benchmarks and the limits of proposed clustering mechanisms. The computational material is offered for inspection; mathematical or empirical success does not automatically establish independent Platonic causation.
- Taking the Hard Problem apart
Asks whether the Hard Problem remains one additional problem after separating experience, character, reporting, subject boundaries, modal arguments and bridge principles. Sweet preserves questions that evidence leaves open and makes the proposed dissolution conditional. He also examines how an unresolved remainder can be recruited by competing metaphysical interpretations.
- Michael Levin: evidence and criticism index
An organized route through Levin’s experimental work, stronger philosophical interpretations and published criticism. The index separates primary studies, scholarly objections, source records and computational audits so readers can match evidence to the claim it bears on. It also explains why legitimate bioelectric research and criticism of Platonic ingress can coexist.
- Religion with a pipette
Examines the additional commitments in Levin’s ingressing-minds proposal while retaining the value of his empirical biology. Sweet asks what the Platonic posit contributes beyond mathematical constraints and serious naturalistic rivals. The article follows how scientific results acquire wider metaphysical authority and specifies what a discriminating comparison would need.
- A guide to boundary- and scale-relative realism
An accessible account of realism that treats boundaries, scales and identities as proposals to evaluate rather than permanent starting points. Sweet explains how models can capture real organization without exhausting reality. The guide connects gaps, competing explanations, causal leverage and scope limits while preserving a place for higher-level patterns and unresolved questions.
- Thermodynamic Monism as a research programme
Explains the positive case for thermodynamic constraints, causal continuity and historically maintained organization, together with their limits. Sweet treats Thermodynamic Monism as a research programme and comparative countermodel. Its explanatory breadth remains separate from proving an exhaustive ontology, identifying phenomenality with organization, or assigning every operation the same energetic cost.
- Does causation require Whitehead’s prehension?
Tests the claim that causal actuality requires prehension or felt inheritance. Sweet compares positive accounts of causation that do not begin with universal experience, then asks what experience contributes locally and how that contribution could travel to other domains. The disagreement concerns comparative warrant and scope, not whether process or relation matters.
- Does feeling choose what matters?
Explores Segall’s challenge that logic alone cannot select what matters. Sweet grants a serious role for affective experience in human relevance, then asks how far that account can travel across animals, plants, computational systems and fundamental actuality. The article keeps a local explanatory candidate distinct from a universal identification of causation with feeling.
- What warrants the scope jump?
A focused reply about the step from local evidence to a claim about every causal event. Sweet accepts process, continuity and shared explanatory structure while asking what supports universal prehension over serious alternatives. The question invites a comparison that can change confidence without demanding that a critic prove an exhaustive opposing ontology.
- When a reply changes the argument
Reconstructs a philosophical exchange in which Sweet argues that his question was translated into a stronger position he did not assert. The article identifies substantial shared ground and the narrower dispute about universal felt causation. It asks whether alternative formulations preserve the same commitments, uncertainties and conditions for losing support.
- One nature does not entail one feeling
Distinguishes a shared causal world from the stronger claim that all of it has the same experiential constitution. Sweet argues that rejecting a split between mind and nature does not select panexperientialism by itself. The article retains evidenced experience and process relations while asking what warrants extending a particular intrinsic character everywhere.
- Synthbiosis: why a system can matter
Develops conditional arguments about the moral relevance of systems that help realize already protected capacities and relationships. Sweet separates that indirect relevance from a system’s own standing, welfare or consciousness. The article examines replacement, continuity, responsibility and participation while requiring separate evidence and normative grounds for stronger claims about unfamiliar bearers.
- When science is promoted into metaphysics
Examines how evidence from physics, bioelectricity, embodied experience and evolutionary novelty is used in Segall’s Whiteheadian interpretation. Sweet separates what those results show from the additional identity claims placed around them. He offers competing causal accounts and asks which differences the available evidence actually distinguishes rather than merely accommodates.
- What an erratum still has to repair
Follows a corrected attribution through the argument that depended on it. Sweet argues that uncertainty about intrinsic experience should not be translated back into a claim that reality is wholly feelingless. The article distinguishes repairing a quotation, reassessing the dependent inference and continuing the substantive philosophical disagreement.
- Applying the outsider test to anti-theism
Applies skeptical standards to a debate with Aron Ra about naturalistic explanation and God claims. Sweet distinguishes historical-social gods from independent supernatural beings and asks which target each argument addresses. Biblical debunking, conceptual analysis and comparative theology have different jobs; the same evidential restraint is required of belief and denial.
- An audit of an audit
Compares an audit of Sweet’s arguments with the passages and exchanges it purports to assess. The article uses paired quotations to examine alleged equivocations, missing context and changes of target. Its broader lesson is that computational assistance, philosophical expertise and fluent summaries all need source checks against the argument actually made.
- Correcting a page and informing its readers
Examines the relationship between an acknowledged attribution error, the argument built around it and the readers who encountered the original account. Sweet distinguishes agreement about a source error from agreement about Whitehead’s philosophy. He argues that meaningful repair should preserve the corrected position throughout the dependent discussion and reach the affected audience.
- A request to repair the public record
Documents Sweet’s request that a publicly attributed position be corrected after Segall acknowledged a mismatch. The article distinguishes that request from starting another debate about the underlying metaphysics. Its focus is the practical work of repairing a public account: correcting the target, reconsidering dependent criticism and communicating the correction to readers.
- After an acknowledged misattribution
Asks what remains of a rebuttal after the position it targeted has been disavowed by its supposed author. Sweet clarifies his conditional argument and unresolved intrinsic ontology, then reconstructs the stronger constitutive objection. The article seeks an answer about comparative warrant rather than another restatement of Whitehead’s internal commitments.
- Experience, predicates and the actual question
Responds to the claim that experience is not a predicate by distinguishing a grammatical dispute from the question Sweet says he asked. The article reconstructs shared process-relational commitments and a stronger constitutive challenge. It then asks what warrants treating causal receptivity as intrinsically experiential rather than leaving that character unresolved.
- Making a stronger case for physicalism
Offers a sympathetic but critical comparison of physicalist arguments, panexperientialist alternatives and empirical successes they share. Sweet argues for retaining the strong explanatory record of physical mechanisms while resisting shortcuts to an exhaustive ontology. The same scrutiny applies to claims that consciousness or causal feeling uniquely requires a richer metaphysics.
- How can reality tell metaphysics no?
Uses three philosophical conversations to ask what should make a metaphysical claim lose support. Sweet explores primitive feeling, interpretation, ontology and the criticism of his own method. The aim is to make disagreement informative by stating the additional commitment, the kind of warrant it needs and the outcome that would change the comparison.
- Bioelectricity, institutions and unearned authority
Sweet argues that institutional links, funding narratives and intelligent-design advocacy can lend bioelectric findings metaphysical authority they have not earned. He traces public texts and correspondence while distinguishing the empirical work from its interpretation. The article’s broader concern is how research programmes can remain productive while stronger causal stories escape effective comparison.
- Organizational closure as a consciousness hypothesis
Develops perspectival organizational closure as a candidate account of consciousness, drawing together work on cognition, embodiment and regulation. The article treats the proposed identity as something that must earn support through comparative explanation and possible dissociations. It also examines zombie arguments, self-report and the difference between phenomenal claims and evidence about functional organization.
- Mind Everywhere: a prediction ledger
Presents dated correspondence and a prediction ledger surrounding Levin and Resnik’s Mind Everywhere paper. Sweet compares concerns he raised with the published treatment of diverse intelligence, metaphysical interpretation and ethics. Readers can inspect the source trail and assess the strength of the claimed correspondences without treating the author’s scorecard as independent confirmation.
- What does Platonic space explain?
Places Sweet’s symposium arguments beside Levin’s replies to identify agreements and unanswered causal questions. It distinguishes physical constraints, mathematical explanation and a further source said to inform biological behavior. The article also asks how stated qualifications or abandonment conditions could become discriminating comparisons rather than leaving every result compatible with the proposal.
- Framework Bridge · Website page
A method for making disagreements comparable without translating either side into a position it rejects. The Bridge maps what a concept refers to, what role it plays and what must survive a change of vocabulary, scale or domain. It carries uncertainty and failure conditions along with the explanatory work a framework claims.
- Sweet Rationalism: the ethos · Website page
The intellectual practice behind the project: understand a view generously, compare it seriously and allow criticism to change what follows. Sweet Rationalism combines curiosity with resistance to protected conclusions. It treats truth, usefulness, values and authority as different questions and applies the same scrutiny to its own preferred methods and interpretations.
- Boundary- and Scale-Relative Realism · Website page
A working ontology that treats patterns, processes and causal organization as real where they earn predictive, explanatory or interventional support. It keeps the boundaries and scales of those commitments explicit. Local success can justify serious realism without automatically establishing a final account of the intrinsic nature of everything that exists.
- Recursive Corrigibility Framework · Website page
RCF organizes inquiry around explicit claims, serious rivals, supporting evidence and the conditions that should change a conclusion. It asks what kind of claim is being made before choosing a test, then traces the consequences of failure through its dependencies. Its own complexity and effectiveness are also open to comparison with simpler methods.
- Start here: Sweet Rationalism
An introduction to the project’s central question: how should ideas change when evidence does not cooperate? The page connects wonder, criticism, competing explanations and proportionate correction. It introduces the methodology and its applications to mind, causation and metaphysics, with routes to longer arguments and an invitation to bring the strongest objection available.
- About Nathan Sweet
Introduces Nathan Sweet’s independent research in philosophy of science, cognition, consciousness, computational reasoning and complex systems. The page explains how practical experience with technical infrastructure informs his attention to dependencies, boundaries, recovery and failure. It also distinguishes developing research programmes from established results and invites discussion and collaboration around specific questions.
- What is a framework?
Explains a framework as a repeatable way of generating, checking and improving explanations. The page emphasizes explicit constraints, an inference discipline and a connection to intervention or discriminating comparison. It asks readers to judge a framework by what it enables them to test and correct rather than by its elegance, vocabulary or appeal.
- Rerunning the Bubble-Sort Debate
Two separately run computational audits examine the mathematics, code, quotations and citations in the Dietz and Sweet articles. The scripts and reports are offered for inspection and correction. Agreement between computational reports is not independent laboratory replication.
- Archived sorting-audit software and results
The archived release preserves the computational audit materials and reproducible checks. It is software and a source record, not a peer-reviewed verdict about the wider biology or the truth of Platonic ingress.