# Apple Inc. v. Motorola, Inc. > When "Means" Matters: Three Lessons in Patent Law **By Paul Roberts** | 🔍 Critical Spotlight 📊 **3,100 words** | ⏱️ **13 min read** #Case_Brief #35_USC_112f #Critical_Spotlight #Software_Patents #Patent_Law #Daubert January 2, 2026 --- ## When "Means" Matters: Three Lessons in Patent Law **Claims without the word 'means' get a strong presumption against § 112(f). Apple's claims survived because they described _how_ the modules worked, not just _what_ they accomplished**. Apple and Motorola sued each other over smartphone patents—because apparently that's what tech companies do instead of innovating. Apple asserted three patents. Motorola countered with six. Judge Posner, sitting by designation in the Western District of Wisconsin, looked at this mess and granted summary judgment that neither side could get an injunction. Then he dismissed the entire case without a trial. Both sides appealed. The Federal Circuit had to sort through three major issues: (1) whether Apple's claims using the word "module" invoke § 112, ¶ 6 means-plus-function claiming, (2) whether Apple's damages expert testimony should have been excluded, and (3) whether either side could get an injunction. **Translation:** The court needed to decide if Apple's patent claims were too vague, if Apple's damages calculations were too speculative, and if anyone deserved to shut down the other's smartphone sales. --- ## The § 112, ¶ 6 Question: What Does "Module" Actually Mean? ### The Statute Says What It Says Section 112, ¶ 6 applies when claims recite "means" for performing a function. When that magic word appears, the claim gets limited to the specific structures disclosed in the specification, plus equivalents. Congress put "means" in the statute for a reason. **Translation:** If you use the word "means," you're stuck with whatever you actually described in your patent. You don't get to claim every possible way to do something. Apple's claims used "module" instead of "means." That creates a presumption *against* applying § 112, ¶ 6. The question: Does "module" by itself connote sufficiently definite structure to overcome that presumption? ### The District Court's Analysis The district court found that "module" doesn't connote structure. The court reasoned that because the specification disclosed *all possible* structures for performing the recited functions, the claim term must be purely functional. Therefore: means-plus-function claiming applies. With respect to Judge Posner, that's backwards. Section 112, ¶ 6 asks whether the claim term itself connotes structure, not whether the specification happens to be comprehensive. The two questions are entirely different. ### What the Federal Circuit Got Right The Federal Circuit reversed. Here's why the reversal was correct: **First,** the presumption against § 112, ¶ 6 is strong when "means" is absent. *Lighting World, Inc. v. Birchwood Lighting, Inc.*, 382 F.3d 1354, 1358 (Fed. Cir. 2004). To overcome it, you need to show the term is "a nonce word or a verbal construct that is not recognized as the name of structure" and "is simply a generic placeholder." *Inventio AG v. ThyssenKrupp Elevator Ams. Corp.*, 649 F.3d 1350, 1357 (Fed. Cir. 2011). **Second,** "module" to a software engineer means something specific: a discrete unit of code with defined inputs, processing, and outputs. Think of it as the API versus just describing what the function accomplishes. The claim limitations here specified *how* the modules work—what data they receive, what processing they perform, what they output. **Third,** the specification's comprehensiveness is irrelevant to whether the claim term connotes structure. The district court conflated two separate inquiries: (1) Does the claim term itself denote structure? and (2) If § 112, ¶ 6 applies, what is the corresponding structure in the specification? Those are different questions with different answers. ### The "Heuristic" Example Take Apple's "heuristic for determining a finger/thumb twist event" limitation. This isn't vague handwaving about "appropriate programming." It's a specific algorithm: If the finger angle is less than 27° from vertical, scroll up. If greater than 27°, translate diagonally. **Translation:** That's structure, even though it's code rather than gears. It's like saying "use bubble sort" versus just saying "sort the data somehow." One tells you *how*, the other just tells you *what*. The Federal Circuit held that these limitations connote sufficiently definite structure. The claims avoid § 112, ¶ 6. --- ## The Damages Expert Exclusion: Did the District Court Confuse Daubert with the Merits? ### What Dr. Napper Did Apple's expert, Dr. Napper, used a "hedonic regression" model—econometric analysis that isolates how much consumers value specific features. He analyzed HTC phone sales to determine what premium consumers paid for features similar to those in Apple's patents. The district court excluded this testimony under *Daubert v. Merrell Dow Pharmaceuticals, Inc.*, 509 U.S. 579 (1993), finding the methodology unreliable. ### Where the Analysis Went Wrong *Daubert* gatekeeping focuses on methodology, not conclusions. The question isn't whether the expert reached the right answer—that's for the jury. The question is whether the expert used a reliable method to reach *an* answer. The district court found Dr. Napper's methodology flawed because: (1) he didn't show the HTC features were "substantially similar" to Apple's patented features, and (2) he didn't prove the features drove consumer demand. But wait—those are merits questions, not methodology questions. Whether the features are sufficiently similar and whether they drive demand are factual disputes for the jury. The expert's job is to use reliable methods to quantify value *assuming* the features are comparable. **Translation:** The judge decided the expert's homework was wrong. But *Daubert* only asks whether the expert showed their work using accepted methods. Whether the homework is actually correct is for the jury to decide. ### The Federal Circuit's Reversal The Federal Circuit correctly held that hedonic regression is a generally accepted methodology in economics. The fact that Motorola's experts disagree with Dr. Napper's application doesn't make the methodology unreliable—it makes it contested. Contested evidence goes to the jury, not out the window. **Here's the bottom line:** Expert testimony should be excluded when the methodology is junk science. It shouldn't be excluded because opposing experts disagree about the facts. --- ## The Injunction Questions: When Can You Shut Down Your Competitor? ### The *eBay* Framework *eBay Inc. v. MercExchange, L.L.C.*, 547 U.S. 388 (2006), established that patent injunctions aren't automatic. You need to show: 1. Irreparable harm 2. Inadequacy of monetary damages 3. Balance of hardships favors you 4. Public interest supports an injunction Both Apple and Motorola failed to satisfy these factors. ### Apple's Problem: Proving Causation Apple needed to show the patented features *caused* irreparable harm—that consumers bought Motorola phones specifically because of the infringing features. Apple's evidence didn't establish this causal connection. The Federal Circuit majority found genuine factual disputes about whether the features drove consumer demand. Judge Prost dissented, arguing Apple's evidence was too speculative. Either way, Apple didn't meet the *eBay* standard. ### Motorola's Problem: FRAND Commitments Motorola's '898 patent was subject to a FRAND commitment (Fair, Reasonable, and Non-Discriminatory licensing). When you promise standards bodies you'll license your patents on reasonable terms, you've already agreed that money damages are adequate. **Translation:** You can't promise to license on reasonable terms, then turn around and demand an injunction. That's like a subscription service that says "you can cancel anytime" but then refuses to let you cancel. The Federal Circuit majority suggested an injunction *might* be appropriate if an alleged infringer "unilaterally refuses a FRAND royalty or unreasonably delays negotiations." Judge Prost correctly disagreed in her concurrence. An alleged infringer is entitled to challenge patent validity before paying license fees. Punishing them for doing so makes no sense. --- ## What This Means for Patent Prosecutors **On § 112, ¶ 6 avoidance:** If you want to avoid means-plus-function treatment for software claims, don't just avoid the word "means." Describe your algorithm's inputs, processing, and outputs. Show the *how*, not just the *what*. Apple did exactly that, and the claims survived. **On damages:** Your damages experts need to use accepted methodologies, but they don't need to prove every factual element at the expert qualification stage. Save those disputes for trial. The district court's exclusion here conflated gatekeeping with merits adjudication. **On injunctions:** Standard-essential patents with FRAND commitments are nearly impossible to enforce through injunctions. The remedy is damages, not shutting down your competitor's products. If you want injunctive relief, don't make FRAND commitments. --- ## The Bottom Line Patent claims mean what they say. When drafters avoid "means," courts should respect that choice—assuming the claim terms actually denote structure to someone skilled in the art. Expert testimony survives *Daubert* when the methodology is sound, even if the application is disputed. Disagreement about facts isn't grounds for exclusion. And FRAND commitments mean what they say. You can't promise reasonable licensing, then seek to shut down your competitor's business. The Federal Circuit got all three issues right. --- **Case:** *Apple Inc. v. Motorola, Inc.*, 757 F.3d 1286 (Fed. Cir. 2014). **Key Holdings:** 1. "Module" with algorithmic detail avoids § 112, ¶ 6 2. Hedonic regression survives *Daubert* despite factual disputes 3. FRAND commitments preclude injunctive relief in most cases **Votes:** - Majority (Reyna, J.): Reversed on § 112(f) and damages expert; affirmed on injunctions - Rader, C.J. (dissenting in part): Would have affirmed on Apple's injunction denial - Prost, J. (concurring in part, dissenting in part): Would have affirmed on Apple's damages expert exclusion and injunction denial