They Say I Say Moves That Matter In Academic Writing

What Actually Happens When You Start an Academic Paragraph

You open your literature review and stare at a blank screen. The paper you're summarizing is interesting, but when you write "Smith argues that climate change is accelerating," something feels off. Your advisor says your writing lacks voice, or worse, sounds like a patchwork report. The issue usually isn't that you can't summarize sources. It's that you haven't connected those summaries to your own argument, and your reader has no idea why they should care about what Smith said. This is where the they say / I say pattern becomes useful, not as a decorative formula but as a structural tool for making your claims legible inside an ongoing conversation. Graff and Birkenstein formalized this in their book, but the underlying mechanism predates it by decades. Every academic field runs on a basic premise: scholars read each other's work and respond to it. Your job as a writer is to make that response visible. Without it, you're just producing information deposits rather than participating in a discussion.

They Say I Say Moves That Matter In Academic Writing

The framework breaks down into a few practical moves. The first is laying out what "they" believe. This doesn't mean everyone everywhere, just the relevant voices in your specific subfield. The second move is stating your response—agreement, disagreement, qualification, or extension. The bridge between these two is what determines whether your writing reads like a literature review or an argument. I spent three semesters helping graduate students restructure their thesis chapters because their committees kept returning drafts that felt descriptively thin. What we discovered was consistent: students were summarizing sources paragraph after paragraph without signaling their own position until the end, sometimes not until the conclusion. By the time they got there, the committee had already lost the thread. We started fixing this by inserting the I say component every two or three source summaries instead of saving it for later. It changed the entire document's architecture. The most common template beginners use is the contrast pivot: they say X, but I say Y. It works because it forces a relationship between sources and your contribution. The problem is that students overuse it until every paragraph sounds identical. You'll see thesauri full of variations—nevertheless, however, although, yet—but the mechanical substitution doesn't fix the structural issue. The real question is whether your counter-claim actually engages with the source you're responding to or whether it's just sitting next to it.

Here's the edge case that trips everyone up. You're working with a source whose position is ambiguous, conflicting, or poorly articulated. I ran into this writing my own dissertation when I was trying to engage with a methodological paper that made three separate claims about the same dataset, and those claims contradicted each other across different sections. A simple they say/I say couldn't capture that. What I ended up doing was mapping the internal tension first—showing where the author's own claims conflicted—then positioning my response around that fault line. It took longer than a standard pivot but it was honest about what the source actually contained. Don't force a clean they say/I say frame onto a source that doesn't support one. Sometimes the better move is to acknowledge the ambiguity before you disagree. Another move that matters more than beginners realize is the concession pivot. They say X, and I agree, but I would add Y. This is the one that separates students who can negotiate complexity from those who default to opposition. Your reader doesn't need you to disagree with everything. In fact, strategic agreement makes your eventual disagreements carry more weight because you've demonstrated you actually read the source. Peer reviewers spot blanket disagreement patterns quickly, and it reads as either laziness or immaturity in the field. There's a subtler version worth mentioning: the echo move. You borrow a key phrase from a source and turn it against itself or extend it beyond its original scope. This requires close reading but produces one of the most satisfying rhetorical effects in academic prose. It signals that you're not just reading to refute but that you're genuinely working inside the source's conceptual framework. I used this technique in a journal submission where I took a methodological term from an epidemiology paper and applied it to organizational behavior. The editor's feedback specifically called out the originality of that move, and it came from nothing more than paying attention to what a source actually said rather than treating it as background material.

The biggest mistake I see is treating this as a sentence-level exercise when it's fundamentally a paragraph-level one. Students will write a correct pivot sentence in isolation and then fill the rest of the paragraph with summary. The I say part disappears after the first line. What you're building is a sustained engagement, not a single comparative clause. Every source paragraph should answer: what am I doing with this evidence, not just what does this evidence say. This framework also breaks down in certain contexts. If you're writing a methods section, a results section, or a systematic review with strict reporting requirements, the they say/I say structure has limited utility. You don't need to position yourself against every cited study when you're describing procedure or presenting data. Forcing it into those sections produces awkward, overwrought prose that reviewers find tiresome. Use the pattern where argument lives—in introductions, discussion sections, and literature reviews that synthesize rather than catalog. Don't apply it universally. The other limitation is that fields vary considerably in how much they expect direct engagement with prior work. STEM disciplines sometimes prioritize brevity and directness over extended conversational framing. Humanities and social sciences tend to reward more elaborate positioning. Adjust the density of your they say/I say moves to match disciplinary expectations, not the other way around. A physics paper that spends three paragraphs establishing conversation turns with prior researchers will be seen as padding, not rigor.

To implement this practically, start with your sources and ask two questions before you write any summary: what does this author claim, and what is my relationship to that claim? Write the answer to the second question in one sentence. Then build the paragraph around that relationship. If you can't answer it clearly, you probably haven't read closely enough to engage the source meaningfully, and no amount of pivot language will fix that. Downloadable templates exist online, but they're mostly useful for beginners who need a scaffold. Once you internalize the pattern, you'll stop thinking about pivot words and start thinking about intellectual relationships. That shift is what separates mechanical writing from argumentative writing. The pattern isn't the point. The point is making your reader aware that something is being said, something else is being said in response, and the response matters.