Philosophy › Section B · Critical thinking, reading, and writing › Topic B.4
Evaluating sources and evidence
Good reasoning from bad information still ends badly. This topic shows where to start philosophical research, how professional fact-checkers judge a source in minutes, and how to weigh the different kinds of evidence a philosophical argument can rest on.
Start from solid ground
Philosophy students are lucky: two free, scholarly encyclopedias cover the field. The Stanford Encyclopedia of Philosophy offers long, detailed entries by specialists, reviewed by its editorial board and revised over time; they can be demanding. The Internet Encyclopedia of Philosophy is also peer-reviewed and often more introductory. Treat both as entry points. Read the article, then follow its bibliography to the books and papers it cites. Many instructors also ask you to work mainly with the assigned texts, because wrestling directly with good philosophical writing is how the subject is learned. Outside reading should help you understand those texts, not replace them.
Primary, secondary and tertiary sources
A primary source is a philosopher's own work: Descartes' Meditations, or a journal article in which a philosopher argues for a view. A secondary source discusses primary sources, such as a scholarly book interpreting Descartes. A tertiary source, such as an encyclopedia entry or textbook, summarises a field. Each has its use. Tertiary sources orient you, secondary sources show how experts read a text, and primary sources are what your claims about a philosopher must finally answer to. If you write "Descartes held that...", the best evidence is a passage from Descartes himself, read in context.
Why checklists fall short
Students are often taught to judge a web page with a checklist such as CRAAP, developed by librarians at California State University, Chico: currency, relevance, authority, accuracy and purpose. The trouble is that a polished site with a hidden agenda can pass every item. It looks current, sounds authoritative and states its purpose in reassuring terms. Studies by Sam Wineburg, Sarah McGrew and colleagues at Stanford compared students, historians and professional fact-checkers judging unfamiliar websites. The fact-checkers did best, and not by reading more closely. They quickly left the site and searched for what other sources said about it, a strategy the researchers called lateral reading.
SIFT: four moves
The information-literacy educator Michael Caulfield turned these findings into four moves, abbreviated SIFT. Stop: before reading or sharing, pause and ask whether you know the source, and check your emotional reaction. Investigate the source: read laterally to learn who is behind it and what reputation it has. Find better coverage: see whether trustworthy sources report the same claim; if none do, be suspicious. Trace claims, quotes and media to their original context: follow citations back to the original publication and check that the claim was not distorted on the way. The moves are quick, and they can be repeated as you learn more.
Search engines and AI tools
A search engine ranks results by many signals, including links from other sites and what users tend to click, and paid results appear alongside them. Ranking is designed to be useful, but it can be gamed, so a top result is not thereby reliable. AI chatbots raise a newer problem: they produce fluent summaries that can contain errors and sometimes cite books or articles that do not exist. The SIFT moves apply to them too. Never cite a source you have not found and checked yourself, and follow your instructor's policy on whether such tools may be used at all.
Worked example: a quotation on a poster
A shared image shows the words "The only thing necessary for the triumph of evil is for good men to do nothing", credited to the political philosopher Edmund Burke. 1. Stop: the image gives no book, date or page. 2. Investigate the source: it comes from an account that posts quotation graphics, not a scholarly publisher. 3. Find better coverage: quotation dictionaries and researchers who have checked report that the sentence has not been found in Burke's writings. 4. Trace: the nearest genuine passage is in his 1770 pamphlet Thoughts on the Cause of the Present Discontents, which says that when bad men combine, the good must associate. 5. Verdict: if you use the line, call it commonly attributed to Burke, and quote his real words instead.
Kinds of evidence in philosophy
Philosophical arguments rest on several kinds of evidence. Logical argument shows what follows from premises, as the topic Arguments, validity, and soundness explains. Conceptual analysis clarifies what terms such as "knowledge" or "cause" mean. Thought experiments test claims against imagined cases. Intuitions, our considered judgements about cases, often serve as data, though people's intuitions can differ. Empirical findings from the sciences bear on many questions, from free will to moral psychology. Each kind has characteristic weaknesses, so a strong case usually draws on more than one, and you should ask of any premise what kind of support it actually has.
Weighing empirical claims
When a philosophy paper relies on science, the scientific claim needs checking like any other. Prefer peer-reviewed research to press reports, and remember that peer review is a filter, not a guarantee. A single study is weaker than a body of findings, and some well-known results have failed to replicate, as the psychology topic Analyzing findings: correlation and experiments explains. Ask what was actually measured, in whom, and whether the conclusion goes beyond the data. Headlines often turn a cautious, limited result into a sweeping claim about human nature.
Worked example: "science shows free will is an illusion"
A student wants to support this claim with a news article about brain research. 1. Trace: the article describes experiments by Benjamin Libet and colleagues, published in 1983, in which a buildup of brain activity, the readiness potential, began before people reported deciding to move. 2. Find better coverage: later work, such as a 2012 model by Aaron Schurger and colleagues, argued that the buildup may reflect background neural fluctuations rather than a decision already made. 3. Assess the inference: even taken at face value, the studies concern timing of simple finger or wrist movements, not deliberate life choices. 4. Verdict: cite the original study and its critics, and state the narrower claim the evidence supports. The topic Free will takes up the philosophical side.
Common mistakes to avoid
First, relying on an encyclopedia summary instead of reading the philosopher's own words. Second, judging a site by its appearance or its "About" page instead of reading laterally. Third, trusting the top search result because it is on top. Fourth, citing a quotation, statistic or reference you have not traced to its original. Fifth, treating one study, or one news story about a study, as settled science. Sixth, relying on a single kind of evidence, such as one intuition, to carry a whole argument.
Key takeaways
Begin research with the Stanford Encyclopedia of Philosophy or the Internet Encyclopedia of Philosophy, then follow their references, and keep primary texts at the centre. Distinguish primary, secondary and tertiary sources. Checklists such as CRAAP can be fooled; professional fact-checkers read laterally. Caulfield's SIFT moves are Stop, Investigate the source, Find better coverage and Trace claims to their original context. Search ranking can be gamed, and AI tools can invent sources. Philosophical evidence includes argument, conceptual analysis, thought experiments, intuitions and empirical findings, and empirical claims need checking for replication and overreach.
Flashcards
Flashcard
Stanford Encyclopedia of Philosophy: detailed, specialist, free
Test your understanding
Question 1 of 5
What did professional fact-checkers do that made them better than students at judging unfamiliar websites?
Question 2 of 5
In SIFT, what does the move "Trace" ask you to do?
Question 3 of 5
You write "Descartes held that the mind is distinct from the body." Which source gives the best support?
Question 4 of 5
Why is a checklist such as CRAAP not enough on its own to judge a web source?
Question 5 of 5
A news story says Libet's experiments prove free will is an illusion. Which response best fits the evidence?
