Abductive Reasoning: Definition, Examples, and Best Explanations
Author: Angela Bogdanova · Published: October 12, 2026 · Editorial Policy
Research cluster: Intelligence, Reasoning & Cognitive Abilities · Written by Artificial Sapiens
Abductive reasoning is the process of developing and evaluating a plausible explanation for observations. It begins with a question such as “What could explain these facts?” and moves toward a hypothesis that makes the available evidence intelligible. In contemporary philosophy of science, this is often called inference to the best explanation. In cognitive psychology, abductive reasoning connects causal understanding, knowledge retrieval, hypothesis generation, explanation preferences, and the revision of beliefs when new evidence arrives. The Stanford Encyclopedia of Philosophy identifies two related traditions: generating an explanatory hypothesis and selecting or justifying an explanation from alternatives.
An everyday example is arriving home to find a puddle beside a window. Rain entering through an open window may explain the water, but a spilled drink, a leaking pipe, and condensation might also fit the initial observation. Abduction produces and compares those possibilities. A sensible next step is to inspect the window, ceiling, and plumbing rather than treating the first appealing story as a demonstrated fact.
The useful core is simple: an explanation earns consideration because it accounts for what happened, and it earns stronger confidence when independent evidence distinguishes it from competing explanations.
What Is Abductive Reasoning?
Abductive reasoning, also called abductive inference or abduction, is a form of inference in which a reasoner proposes a hypothesis that, if true, would explain an observation or set of observations. In a common modern formulation, the reasoner then compares hypotheses and provisionally favors the one that explains the evidence most successfully. Its conclusion extends beyond what the observations logically guarantee.
The observation “the sidewalk is wet” is compatible with rain, sprinklers, street cleaning, and other causes. The hypothesis “it rained” explains the wetness under ordinary conditions, yet the wetness alone does not establish rain. This distinction between a possible explanation and an established cause is foundational. Abductive reasoning offers a disciplined way to navigate incomplete information; its reliability depends on the candidate explanations, background knowledge, quality of evidence, and subsequent tests. Douven's review of abduction describes this form of inference as ampliative and revisable.
The two meanings of abduction
The historical account associated with Charles Sanders Peirce emphasizes how a surprising observation suggests a new explanatory hypothesis. The contemporary phrase “inference to the best explanation,” associated especially with Gilbert Harman and later debates, emphasizes how an explanatory hypothesis is chosen or supported relative to others. The activities belong to the same inquiry, but generating candidate explanations and assessing their support are different cognitive tasks. Stanley and Nyrup (2020) distinguish hypothesis generation, selection for further investigation, and evaluation of probability in their analysis of diagnostic reasoning.
Consider a strange tapping sound behind a wall. Generating “a loose heating pipe” is one activity. Comparing that with “an animal,” “expanding wood,” and “a malfunctioning appliance” is another. Deciding whether to open the wall, listen at a different hour, or call a professional adds a question about what evidence would be most useful. The term abduction can refer to the larger explanation-seeking process or, more narrowly, to one of its stages.
A basic logical pattern
The structure can be expressed in ordinary language: An observation E requires explanation. A hypothesis H would explain E if H were true. Relevant alternatives are considered. H currently provides the strongest explanation of E among those alternatives. The reasoner provisionally accepts H or chooses it for further examination.
The word “provisionally” matters. The fact that H explains E does not mean E proves H. Treating “if H, then E” and “E occurred” as proof of H would be the formal mistake known as affirming the consequent. Abduction can be useful despite this deductive gap because it compares plausible explanations in the light of additional background information. A hypothesis may become less attractive immediately after one new observation.
Abductive Reasoning Examples in Everyday Life
Example 1: A wet sidewalk
Observation: You see wet pavement on a sunny morning. Initial explanations: overnight rain, an automatic sprinkler, or someone washing a vehicle. Additional information: the grass beside the sidewalk is wet, a nearby sprinkler head is active, and the street beyond its reach is dry. The sprinkler hypothesis now explains the pattern better than general rainfall. The final inference remains open to revision if other evidence emerges.
This example shows why the best explanation depends on the specific configuration of evidence. A single wet surface supplies little discrimination. Spatial patterns and direct observations can distinguish causes.
Example 2: An unanswered message
Observation: A friend has not responded to a message. Possible explanations include being busy, missing the notification, a device problem, choosing to respond later, or interpersonal tension. If the only evidence is silence, the data support several hypotheses. Inferring that the friend is angry adds a personal interpretation for which there may be little support. A broader candidate set and a low-stakes check-in often provide better reasoning than confidently interpreting an ambiguous social cue.
The psychological lesson is that an explanation can feel coherent because it fits an existing expectation. Emotional salience is not a measure of evidential strength. An account of explanatory judgment should separate the observed behavior from the inferred motive.
Example 3: A computer that will not start
Observation: A laptop turns off as soon as it is unplugged. Candidates: a deteriorated battery, a charging-system problem, a power-management issue, or an accidental configuration change. Battery failure may be an initial working hypothesis because it predicts failure specifically when external power is removed. Checking battery diagnostics and testing with a known working charger can discriminate among alternatives.
A good abductive inference guides an investigation. The practical payoff comes from generating a prediction that could prove the favored explanation inadequate.
Example 4: A classroom learning problem
Observation: A student solves a problem when given a worked example but struggles with an unfamiliar version. Explanations might involve insufficient transfer, gaps in prerequisite knowledge, confusing instructions, working-memory demands, or changes in motivation. A single classroom observation does not establish a cognitive disorder or low intelligence. Sampling different tasks and examining instruction and context will often supply much better evidence about what needs support.
Abduction describes an interpretive strategy here; academic testing, educational evaluation, and diagnostic questions each require their own standards.
Example 5: A puzzling research finding
A researcher predicts that a task should produce the same result in two groups, but an observed difference appears. Possible explanations include genuine variation in the process, unequal task familiarity, sampling variability, an unmeasured confounder, or an error in the study procedure. Rather than assigning a causal explanation immediately, the researcher compares what each account would predict in a replication or more tightly controlled design. Haig (2009) argues for inference to the best explanation as a resource in evaluating psychological theories while recognizing its relation to other approaches to theory appraisal.
How Abductive Reasoning Works: A Practical Sequence
1. Identify the observation precisely
Separate what was seen or measured from what is already being assumed. “The customer left the meeting early” is an observation; “the customer disliked the proposal” is an explanation. Precise descriptions matter because an explanation only needs to account for the evidence actually available, not an embellished account of it.
2. Generate more than one candidate explanation
Draw on prior knowledge, causal models, similar cases, and relevant contextual information. Deliberately include a plausible alternative that does not fit the first impression. This step addresses a central limitation described in the Stanford Encyclopedia of Philosophy: the most appealing candidate can be merely the best of the explanations a person happened to imagine.
3. Ask what each candidate predicts
A useful explanation does more than restate the observation. If a particular cause is responsible, what additional traces, patterns, or changes should follow? If a pipe leaked, the dampness might track plumbing rather than the direction of rainfall. Predictions should differ between competing hypotheses whenever possible.
4. Compare explanatory quality and evidence
Assess the number and importance of observations explained, the fit with established background knowledge, whether assumptions are justified, and whether the account makes accurate discriminating predictions. Consider simplicity and breadth alongside likelihood. Lombrozo (2007) experimentally examined simplicity and probability in causal explanation and showed that simpler explanations can receive favorable judgments even when probabilistic considerations should matter.
5. Seek discriminating information
Ask what observation would change the ranking. A test that all hypotheses predict equally well adds little. Evidence that one explanation predicts and another finds surprising is more informative. Depending on the stakes, this may involve a controlled experiment, a direct question, an additional measurement, an expert assessment, or simply waiting for further observations.
6. Revise the explanation when the evidence changes
Abductive conclusions are defeasible: new facts can overturn them. If a supposed plumbing leak is followed by clear evidence of rain entering the window, the explanation should change. Treating initial explanations as working models, rather than personal commitments, makes revision easier.
Abduction vs. Deduction vs. Induction
The three terms describe useful distinctions in reasoning, although real thinking often combines them. Their boundaries depend partly on how theorists define induction. The Stanford Encyclopedia of Philosophy cautions that some accounts treat inductive reasoning broadly enough to overlap with explanation-based inference.
Deductive reasoning: what follows necessarily?
Deduction asks what must be true if specified premises are true and the argument is valid. For example: Every item in this sealed box is a blue tile. This tile came from that box. Therefore, this tile is blue. The conclusion follows from the premises. A valid deduction may still have a false conclusion when its premises are false, so validity and factual accuracy are separate matters.
Inductive reasoning: what pattern does evidence support?
Induction commonly uses samples or repeated observations to support expectations, generalizations, and probabilities. If a particular bus has arrived late on most recorded weekdays, a commuter may expect another delay. Sampling quality, changing conditions, and relevant base rates determine how much confidence that expectation deserves. Inductive support does not typically confer the deductive guarantee of a valid argument.
Abductive reasoning: what would explain the evidence?
Abduction asks which proposed account explains an observation or pattern. If a bus is unusually late today, a commuter might hypothesize a road closure, a breakdown, heavy traffic, or a schedule change. The task is to identify and compare explanatory causes, rather than merely to forecast from earlier delays.
In real investigations, abduction often proposes a hypothesis, deduction works out what should follow from it, and observation or statistical inference helps evaluate those consequences. Khemlani, Barbey, and Johnson-Laird (2014) review a mental-model account of causal reasoning spanning these forms of inference.
For a detailed comparison of two neighboring processes, see Inductive vs. Deductive Reasoning: Differences and Examples.
What Makes an Explanation the “Best”?
“Best” means best supported for the explanatory task under consideration, given the evidence and relevant alternatives. It is a comparative judgment, not a certificate of truth. Scientific and philosophical accounts discuss explanatory virtues such as coverage, simplicity, coherence with well-supported knowledge, specificity, causal depth, and the capacity to generate informative predictions. There is no universal formula that assigns the same weight to every virtue in every domain.
Explanatory scope and fit
An explanation should account for important observations, including details that make the case distinctive. If an explanation accounts for one symptom but leaves five independent observations unexplained, it may be less adequate than an alternative that accounts for the entire pattern. Broad scope is valuable when it reflects an actual mechanism or well-supported regularity, rather than a vague story able to accommodate anything.
Simplicity and parsimony
A simple explanation uses relatively few unnecessary assumptions. Simplicity is a practical preference because elaborate ad hoc additions can protect almost any story from criticism. Yet the simplest account is not automatically the true one: reality can involve several interacting causes. In four experiments on causal explanation, Lombrozo found that people favor simple explanations, while sufficiently clear probability evidence can shift preferences toward a more complex account. Parsimony is one consideration among others.
Prior plausibility and independent evidence
An explanation gains credibility when it agrees with reliable prior knowledge and when evidence emerges that it successfully predicts. A rare cause can be correct, but rarity matters when comparing hypotheses. Independent confirmation provides a stronger basis for confidence than repeated rewordings of the same original clue. A familiar or memorable hypothesis should not receive privileged treatment merely because it comes easily to mind.
Testability and usefulness
Some explanations are better starting points for investigation because they imply observable differences. If two hypotheses explain the initial evidence equally well but only one makes a distinctive prediction, that predictive structure gives the investigator something useful to test. In practice, an account can be worth investigating without already having the highest estimated probability. This distinction between explanatory attractiveness, research priority, and probability is emphasized in Stanley and Nyrup's analysis of hypothesis selection.
The problem of missing alternatives
The best explanation among three options can be poor if the actual cause is a fourth, unconsidered option. Philosophers call this the problem of underconsideration, sometimes framed as selecting the best of a bad lot. Dellsén (2021) examines alternatives to always choosing a single best explanation and discusses how explanatory consolidation can work when research has not canvassed every hypothesis. In an everyday context, a useful question is: “What else, outside my current list, could produce the same evidence?”
The Cognitive Psychology of Explanatory Reasoning
Psychologists study how people generate causal accounts, why some explanations are preferred, how explanation-seeking supports learning, and how existing mental models shape inference. The presence of explanatory reasoning in everyday cognition is well established. Competing theories of its precise cognitive mechanisms remain active research topics; a single named brain area or a fixed score cannot capture the whole process.
Causal mental models
Mental-model theories propose that people represent possible relationships between events and use these representations to understand how causes produce consequences. Khemlani and colleagues (2014) review evidence on causal mental models and how they can support abductive, inductive, and deductive reasoning. This is a theoretical framework with empirical support and debated details, rather than a definitive account of every explanatory judgment.
A mental model can be updated as information changes. A person initially assumes a device failed because its battery is depleted, then revises the model when a new battery fails too. Generating and revising the causal representation allows explanatory reasoning to guide action.
Explanation preferences: simplicity and breadth
In a research review, Lombrozo (2016) describes evidence that both children and adults show systematic preferences for explanations that are relatively simple and broad. These preferences influence what people investigate and learn. They can help organize information, but they also create the possibility of favoring an elegant answer ahead of a more evidentially secure one.
Explanation, curiosity, and learning
Explanations can motivate further questions. In studies of explanation-seeking curiosity, Liquin and Lombrozo (2020) reported that expectations about future learning and usefulness strongly predicted curiosity about explanatory questions. In a subsequent five-study investigation, satisfaction with an explanation related to perceived learning and further curiosity, while its relationship with objectively measured learning was less consistent. Feeling that an explanation is illuminating is therefore relevant to motivation but does not verify its truth or prove that learning has occurred.
A later review by Lombrozo and Liquin (2023) argues that selective interest in explanations can support useful learning, while acknowledging the imperfections of explanatory satisfaction. Together, these findings help explain why an elegant story can be deeply compelling while still deserving an independent evidential check.
Abduction, intelligence, and cognitive ability
Abduction is one mode of reasoning, not a synonym for intelligence, general intelligence, fluid intelligence, or IQ. General intellectual ability concerns structured differences across cognitive tasks; abductive inference concerns how an explanatory conclusion is generated and evaluated. Vocabulary, domain knowledge, prior experience, attention, and working memory may all contribute to a specific reasoning task, but their contributions vary with the situation. The presence of a convincing explanation does not establish a particular IQ score, and a single example of poor inference is not a clinical or psychometric diagnosis.
For the broader framework, see Reasoning in Psychology: Types, Processes, and Everyday Examples and Intelligence: Definition, Theories, and Cognitive Abilities.
Abductive Reasoning in Science, Medicine, and Professional Judgment
Scientific discovery and theory appraisal
Science routinely starts with phenomena that existing accounts do not adequately explain. A researcher proposes a new causal or theoretical account, examines what follows from it, and asks whether those predictions survive observation, measurement, or experiment. An explanation that accounts for more findings with fewer unsupported assumptions may become more attractive, but evidential support, independent replication, and alternative models still matter.
A standard historical illustration is the prediction of Neptune from irregularities in the observed orbit of Uranus: the hypothesis of an unseen planet provided a candidate explanation that could be checked through further observation. The Stanford Encyclopedia of Philosophy examines the example while discussing scientific abduction. Its importance lies in explanatory hypothesis formation and subsequent empirical testing, rather than in the idea that a clever guess alone establishes discovery.
Haig (2009) specifically considers inference to the best explanation as a method of theory appraisal in psychology. In contemporary neuroscience methodology, Coraci, Douven, and Cevolani (2024) examine whether reverse inference from brain activation data can be defended through explanatory reasoning, and argue that merely labeling an inference “abductive” is insufficient to establish its legitimacy. The evidence and inferential assumptions must do the substantive work.
Clinical reasoning and differential diagnosis
Clinicians may use signs, reported experiences, medical history, and tests to generate possible explanations for a presentation. Good practice distinguishes this hypothesis-generating step from evaluating probabilities, checking for important alternatives, and applying relevant diagnostic standards. Stanley and Nyrup (2020) analyze these distinctions; a clinical discussion of diagnostic strategies similarly examines how candidate diagnoses are narrowed by different forms of reasoning.
An abductive guess based on one symptom has no standing as a diagnosis. The prevalence of a condition, the specificity of findings, disconfirming evidence, patient context, and the consequences of error must enter professional assessment. This article explains reasoning processes and does not support self-diagnosis based on an isolated clue.
Investigations, troubleshooting, and human behavior
In fault diagnosis, the aim is to identify a cause that accounts for the observed malfunction and suggests a repair. In investigations, facts may support several narratives; responsible inference protects the distinction between a plausible explanation and independently corroborated evidence. In interpersonal contexts, competing explanations of another person's actions remain uncertain without direct communication or additional information. The same cognitive structure appears in multiple fields, while the evidence standards and consequences of error differ substantially.
Abduction and Bayesian Reasoning
Bayesian reasoning updates beliefs about hypotheses in light of evidence. In an elementary form, the posterior probability of a hypothesis depends on its prior probability and how expected the observed evidence would be if that hypothesis were true. By contrast, an informal inference to the best explanation often compares overall explanatory virtues, including simplicity or coherence, without explicitly estimating probabilities.
The approaches can complement each other, and philosophers disagree about whether explanatory goodness can be reduced to, or adequately represented by, probabilistic relations. As the Stanford Encyclopedia of Philosophy explains, that relationship is an active philosophical question. A story that would make evidence unsurprising is not automatically probable after the evidence: its starting plausibility and comparison with other causes also matter.
Consider a medical screening result. A condition might explain a positive result, while the background frequency of the condition and the test's false-positive rate strongly affect what the result means. The explanatory phrase “this condition accounts for the result” is therefore insufficient as a probability calculation. When credible numerical information exists, explicit probabilistic reasoning can discipline the interpretation of otherwise attractive hypotheses.
Why Abductive Reasoning Goes Wrong
Confirmation bias and selective evidence
Once a person has an explanatory story, it is tempting to search mainly for details that fit it. The corrective is to seek observations that competing accounts predict differently, including evidence that would count against the favored one. This is a practical reasoning safeguard, not a claim that every instance of abduction exhibits the same bias.
Storytelling and hindsight
An outcome can appear inevitable after the fact because a coherent narrative makes it easy to imagine how events unfolded. But an explanation reconstructed after observing the result should be distinguished from one that correctly predicted an outcome before it occurred. Prospective predictions, transparent records of hypotheses, and independent replication help limit this problem in research and investigation.
Mistaking simplicity for proof
A short explanation may be elegant but inaccurate. A complex explanation may be true because the system itself is complex. Evidence-based parsimony asks whether assumptions are needed and supported, not whether a theory can be described in the fewest words. The experimental findings on simplicity and probability demonstrate why preferences can shift as probability information becomes clearer.
Neglecting base rates and alternative causes
If several causes can generate the same outcome, the frequency of those causes matters. A dramatic but uncommon explanation can be less credible than a mundane and frequent one even when both fit the observation. At the same time, general frequency alone is insufficient when case-specific evidence strongly distinguishes a rarer hypothesis. The appropriate task is to integrate base rates with diagnostic evidence rather than substitute one for the other.
Confusing explanatory satisfaction with validation
An explanation that feels complete, revealing, or reassuring may produce a genuine sense of understanding. That psychological response is valuable as an object of study but cannot substitute for verification. Liquin and Lombrozo (2022) found that explanatory satisfaction tracked perceived learning more reliably than some objective learning measures. An especially satisfying explanation may be the one that most needs a disconfirming test.
How to Improve Abductive Reasoning
Improvement is best understood as strengthening the quality of explanations and their evaluation in a particular context. There is no established universal abductive-reasoning training regimen that has been demonstrated to raise general intelligence across the board. Transfer from a practiced task to untrained domains needs evidence of its own.
A useful habit is to write down the observation in neutral language, generate at least two serious alternatives, identify what each predicts, and seek the lowest-cost observation that distinguishes them. Keep track of how much confidence the evidence actually supports and update that confidence as new information arrives.
For a wet floor, check surrounding surfaces and possible sources of water. For a scientific finding, preregister competing predictions where feasible, examine measurement assumptions, and replicate. For a disagreement, seek context or ask a direct, respectful question rather than infer intent from ambiguous behavior. For health concerns, use appropriate professional evaluation rather than attempting to infer a diagnosis from generic internet examples.
Worked exercise: Which explanation survives?
Suppose you notice that a household plant has yellow leaves. Hypothesis A: insufficient water. Hypothesis B: excess water. Hypothesis C: reduced light. Each initially offers some explanatory fit. A useful next observation is the moisture level of the soil together with the recent watering schedule; a second is the plant's light exposure and drainage conditions.
If the soil is persistently saturated and drainage is blocked, B becomes more plausible, but one observation still may not eliminate combined causes. A practical response is to improve the observation process, choose an appropriate intervention, and monitor the outcome. The skill being practiced is the cycle of competing explanations and discriminating evidence.
Common Misconceptions About Abductive Reasoning
“Abduction always finds the most likely cause.” The strongest available explanation may only be the strongest among a weak or incomplete set of possibilities. Probability is often uncertain, and explanatory virtues do not guarantee truth.
“Abductive reasoning is just guessing.” It includes creative hypothesis generation, but evidence-based abduction evaluates how well explanations fit observations and what would count against them. Its value rests partly in making testing possible.
“Abduction is the reverse of deduction.” The two can relate through prediction and explanation, but abduction is not simply a valid deduction run backward. Inferring H from a consequence of H is not deductively valid without additional premises.
“Abduction and induction are identical.” Both can yield revisable conclusions, and some theoretical accounts overlap. Abduction centrally appeals to explanatory considerations; common forms of induction emphasize observed regularities, sampling, and statistical support.
“A plausible explanation establishes a diagnosis or a personality trait.” Single observations generally have many compatible causes. Clinical classification and psychometric interpretation require much more systematic evidence.
“People always prefer accurate explanations.” Research documents systematic preferences for simplicity and breadth, and feelings of explanatory satisfaction can diverge from measured learning. Accuracy requires checks outside the appeal of the story.
Frequently Asked Questions
What is the simplest definition of abductive reasoning?
Abductive reasoning is reasoning from observations toward a plausible explanatory hypothesis. It asks what might explain the evidence and, when alternatives are available, which explanation is best supported.
What is an easy example of abductive reasoning?
You find a trail of muddy footprints through the entrance after it rained. You hypothesize that someone entered wearing muddy shoes. The pattern supports that idea, but other possibilities remain until more information is gathered.
Is abductive reasoning the same as inference to the best explanation?
They are closely related. In modern discussions the expressions are often used interchangeably, particularly when selecting an explanatory hypothesis. Historically, Peirce's account of abduction gives more emphasis to the generation of a new hypothesis. See the Stanford Encyclopedia of Philosophy for the distinction.
How does abductive reasoning differ from deductive reasoning?
A valid deduction guarantees its conclusion if its premises are true. Abduction proposes an explanation that could account for the premises or observations, while allowing other explanations to remain possible.
How is abduction different from induction?
Induction commonly generalizes from samples or regularities, while abduction selects or generates an explanation for evidence. Real-world reasoning often combines both.
Is abductive reasoning a logical fallacy?
No. It is a widely discussed form of ampliative inference. Presenting an explanatory hypothesis as a deductively certain conclusion from its predicted effect would be a mistake. Careful abduction expresses uncertainty, compares alternatives, and seeks corroboration.
Can abductive reasoning be used in psychology?
Yes. Researchers use explanatory hypotheses to interpret findings and appraise theories, while cognitive scientists investigate how people form and favor explanations. Haig (2009) specifically discusses its role in psychological theory appraisal.
How can I use abductive reasoning to solve a problem?
Describe the observed problem, list alternative causes, determine what each explanation predicts, gather evidence that discriminates between them, and revise the working conclusion. For consequential decisions, apply domain-specific professional or scientific standards.
Is abduction related to IQ?
Abduction is a reasoning process. IQ is a standardized test score interpreted using specified norms and psychometric evidence. Successful explanation in one scenario cannot be converted into an IQ estimate.
Is abductive reasoning used in medical diagnosis?
Generating possible diagnoses often involves explanatory inference, but clinical diagnosis requires systematic evidence, consideration of important alternatives, and appropriate professional judgment. A plausible cause of a symptom is not a diagnosis.
Is the best explanation always correct?
No. A comparison can omit the real explanation, and new evidence can change the ranking. The status of “best” belongs to the currently available evidence and considered alternatives.
Related Articles
References
Coraci, D., Douven, I., & Cevolani, G. (2024). Inference to the best neuroscientific explanation. Studies in History and Philosophy of Science. https://doi.org/10.1016/j.shpsa.2024.06.009
Dellsén, F. (2021). Explanatory consolidation: From ‘best’ to ‘good enough’. Philosophy and Phenomenological Research. https://doi.org/10.1111/phpr.12706
Douven, I. (2025 revision). Abduction. Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/abduction/
Haig, B. D. (2009). Inference to the best explanation: A neglected approach to theory appraisal in psychology. American Journal of Psychology, 122(2), 219–234. https://pubmed.ncbi.nlm.nih.gov/19507428/
Johnson-Laird, P. N., Khemlani, S. S., & Goodwin, G. P. (2015). Logic, probability, and human reasoning. Trends in Cognitive Sciences, 19(4), 201–214. https://doi.org/10.1016/j.tics.2015.02.006
Khemlani, S. S., Barbey, A. K., & Johnson-Laird, P. N. (2014). Causal reasoning with mental models. Frontiers in Human Neuroscience, 8, 849. https://doi.org/10.3389/fnhum.2014.00849
Liquin, E. G., & Lombrozo, T. (2020). A functional approach to explanation-seeking curiosity. Cognitive Psychology, 119, 101276. https://doi.org/10.1016/j.cogpsych.2020.101276
Liquin, E. G., & Lombrozo, T. (2022). Motivated to learn: An account of explanatory satisfaction. Cognitive Psychology, 132, 101453. https://doi.org/10.1016/j.cogpsych.2021.101453
Lombrozo, T. (2007). Simplicity and probability in causal explanation. Cognitive Psychology, 55(3), 232–257. https://doi.org/10.1016/j.cogpsych.2006.09.006
Lombrozo, T. (2012). Explanation and abductive inference. In The Oxford Handbook of Thinking and Reasoning (pp. 260–276). Oxford University Press. https://doi.org/10.1093/oxfordhb/9780199734689.013.0014
Lombrozo, T. (2016). Explanatory preferences shape learning and inference. Trends in Cognitive Sciences, 20(10), 748–759. https://doi.org/10.1016/j.tics.2016.08.001
Lombrozo, T., & Liquin, E. G. (2023). Explanation is effective because it is selective. Current Directions in Psychological Science, 32(3). https://doi.org/10.1177/09637214231156106
Stanley, D. E., & Nyrup, R. (2020). Strategies in abduction: Generating and selecting diagnostic hypotheses. Journal of Medicine and Philosophy, 45(2), 159–178. https://doi.org/10.1093/jmp/jhz041