Module 1: Legal Research Fundamentals

Every source you’ll ever cite falls into one of two categories. Primary authority is the law itself — a judicial opinion, an enacted statute, an administrative regulation, a constitutional provision. Secondary authority is commentary about the law — a treatise, a legal encyclopedia entry, a law review article, a Restatement. Secondary sources are never binding on a court, no matter how well-reasoned or widely cited.

Why the distinction matters in practice. A court will never rule for your client because a treatise said so — but a treatise can save you hours by pointing you straight to the primary cases that actually control. Confusing the two — citing a secondary source as if it were itself the law, or skipping secondary sources entirely in unfamiliar territory — is one of the most common mistakes a new researcher makes.

The distinction has taken on new, genuinely urgent stakes in the AI era. In Mata v. Avianca, Inc., 678 F. Supp. 3d 443 (S.D.N.Y. 2023), attorneys representing a personal injury plaintiff filed a brief citing more than half a dozen judicial opinions to support their legal arguments — none of which actually existed. The attorneys had used ChatGPT to help draft the brief, and the tool had confidently generated realistic-looking case names, reporter citations, and even fabricated quoted holdings, none of which the attorneys had independently verified against an actual reporter or database before filing. The court imposed sanctions, specifically emphasizing that the attorneys’ underlying failure wasn’t using an unfamiliar tool — it was never actually confirming, through a real primary source, that the cases they were relying on genuinely existed and said what they claimed. Mata v. Avianca is now the standard cautionary citation for exactly this section’s core point: a citation is only worth what the primary source behind it actually says, and no source — AI-generated or otherwise — substitutes for independently verifying the law itself.

Once you’ve found primary authority, the next question is how much it actually controls. Mandatory (binding) authority comes from a court with real power over the court deciding your case — typically a higher court in the same jurisdiction, on the same issue. Persuasive authority — an out-of-state decision, a lower court’s dicta, a well-reasoned secondary source — may genuinely influence a court’s reasoning, but it cannot compel a particular result.

This isn’t a fixed property of a case in the abstract — it depends entirely on which court is deciding which question. A decision that’s mandatory authority in one court’s courtroom is only persuasive in another’s.

A real, common trap: federal court doesn’t mean federal law. Under the Erie doctrine (from Erie Railroad Co. v. Tompkins, 304 U.S. 64 (1938)), a federal court hearing a state-law claim under diversity jurisdiction must apply that specific state’s own substantive law, not some generalized federal common law. A researcher working a federal diversity case who reflexively pulls federal appellate decisions on a state-law question is chasing authority that may be merely persuasive — or entirely irrelevant — while the actual mandatory authority sits in that state’s own courts. Getting mandatory authority right isn’t simply about court level, it requires first confirming exactly which sovereign’s law actually governs the question being researched.

The digest and key-number system organizes case law by legal topic rather than by court or date — the same standardized key number is assigned to the same specific point of law wherever it appears, letting a researcher pull together cases addressing that exact point across many jurisdictions at once.

An annotated code is a statute supplemented with case-note annotations under each section, showing which courts have interpreted and applied that specific provision. Since statutory language is often open-ended (a “reasonable time,” a “material breach”), the case notes are frequently what actually tells you what the statute means in practice.

The key-number system has a real, specific commercial origin worth knowing. It was created by John B. West and the West Publishing Company, which he founded in St. Paul, Minnesota in 1872. West’s real innovation, the National Reporter System, launched in 1879, was the first system to publish American case law comprehensively and consistently by geographic region, rather than leaving courts and individual publishers to print opinions inconsistently or not at all — before West, finding a case from a different state often meant tracking down a separate, sometimes hard-to-obtain state-specific reporter. The key-number classification system followed as the natural next step: since West was already publishing essentially all American case law in one integrated system, assigning every point of law a standardized topic-and-subtopic code let researchers trace a single legal issue across every jurisdiction West published, all at once. That nineteenth-century commercial innovation is the direct ancestor of the digest research method this section teaches, and West’s key-number taxonomy (now owned by Thomson Reuters) remains in active use today, well over a century later.

Research doesn’t end when you run out of time — it ends when it’s actually done. Research saturation is the genuine signal: new sources keep circling back to the same small set of core authorities you’ve already found, rather than surfacing anything new. That convergence is what tells you the controlling framework has actually been mapped, not just that you’ve gotten tired of searching.

The opposite failure is just as real. Stopping the moment you find one case that seems to support your position — without checking whether it’s mandatory authority, whether it’s still good law, or whether contrary authority exists — isn’t efficient research. It’s an incomplete one that happens to feel finished.

This isn’t just good practice — it’s a real, enforceable professional obligation. ABA Model Rule of Professional Conduct 1.1 requires a lawyer to provide competent representation, and the rule’s own official comment specifically defines competence to include the “thoroughness and preparation reasonably necessary for the representation” — language state disciplinary bodies and malpractice courts have repeatedly read to require genuinely adequate legal research, not merely a quick search that happened to turn up something plausible-looking. A lawyer who stops researching the moment one favorable-looking case surfaces, without confirming it’s actually mandatory authority, still good law, and free of unaddressed contrary authority, isn’t just doing weaker work — they may be falling short of the actual professional competence standard the rules of professional conduct impose. Research saturation, in other words, isn’t a purely stylistic preference for thoroughness; it tracks a real, enforceable ethical floor.