Is my research topic original? How to check before you commit
The worst time to discover your question has been answered is eight months in, with a supervisor meeting on Thursday.
25 August 2026
Almost everyone asks this question too late. You choose a topic because it interests you, you read around it for a few weeks, nothing obviously identical turns up, and you take that silence as permission. Then somewhere in the second year you find the paper. It is usually in a journal you do not read, by someone in a neighbouring discipline, using different words for the same thing.
This is avoidable. Not perfectly, because you cannot prove that something has never been published. But the check that catches most of it takes an afternoon, and almost nobody does it properly, because they are asking the wrong question first.
What originality actually means
The question in your head is probably "has anyone studied this before?" If the answer is yes, you conclude the topic is taken. That is the wrong test, and it is why people abandon perfectly good projects.
Nothing worth researching is untouched. If literally nobody has written about your subject, that is far more often a warning than an opportunity. It usually means the question is not answerable, not interesting, or already settled so long ago that it stopped being discussed.
Originality is not an empty space in the literature. It is a defensible claim that your specific contribution has not been made.
There are several ways to make one, and only the first is what people usually imagine.
- A new question. Nobody has asked this. Rare, and worth being suspicious of.
- A new context or population. The question has been answered somewhere else, for someone else. Whether that counts depends entirely on whether there is a reason to expect a different answer. "Nobody has done it in my country" is not a contribution. "The mechanism depends on a factor that differs in my country" is.
- A new method. The same question, approached in a way that addresses a known weakness in how it was approached before.
- New data. Access to something nobody else had, or could get.
- A resolution. Two bodies of work disagree, and you can explain why. This is often the strongest and the most overlooked.
Before you search for anything, decide which of these you are claiming. The search you run depends on the answer, and a search run without one finds papers about your subject rather than papers about your claim.
The check
Write your claim as one sentence
Not your topic. Your claim. "Adaptive reuse of industrial buildings" is a topic and you cannot test it for originality. "Adaptive reuse reduces embodied carbon more than demolition and rebuild in humid tropical climates" is a claim, and it can be checked.
If you cannot write the sentence yet, that is the actual finding, and it is a useful one. You do not have a project to check.
Read the most recent review article first
Before any keyword searching, find the newest review or meta-analysis in your area and read its conclusions and limitations. A good review tells you the state of the field in an hour, which is faster than assembling it yourself from thirty papers.
Then read its final section. Reviews almost always end by naming what has not been done. That section is written by someone who has read everything you are about to search for.
Search the other vocabulary
This is the step that catches what everyone else misses. Your discipline has a name for your thing. At least one other discipline has a different name for the same thing, and they are not reading your journals either.
Economists call it something else. Engineers call it something else. Public health calls it something else again. Find two or three papers from outside your field that are clearly about the same phenomenon, take the words they use in their titles and keywords, and search again with those. If your idea has already been done, this is where you will find it.
Mine the "future research" sections
Take the eight or ten most relevant papers and read only their closing paragraphs. Researchers routinely end by describing exactly what they would have done with more time or money, which is a list of open questions written by the people best placed to know.
If several independent papers name the same gap, you have found something real. If your idea appears there and nothing has been published on it since, you have both confirmation the question matters and evidence it is still open.
Check the last two years separately
Run your searches again restricted to roughly the last twenty-four months, and include preprints. This is the window that catches people out. Reviews have not absorbed it yet, citation counts are still near zero so ranking buries it, and somebody may be finishing your project right now.
If you find it has been done
First, be pleased. You have found this in an afternoon rather than in your second year, and the paper that scared you is now the most useful thing you have read, because it tells you precisely where the frontier is.
Then look at what it actually did, because a near-identical paper usually leaves room:
- Take its limitations section seriously. The authors have listed the weaknesses of their own work. Addressing one of them is a legitimate contribution and an easy one to defend.
- Check whether anyone replicated it. A single study that nobody has confirmed is not a settled question, particularly if it was small.
- Find the disagreement. If two papers reached different conclusions, explaining why is a better project than either of them.
Knowing when to stop
You cannot prove a negative, and at some point continuing to search becomes a way of avoiding starting. The standard is not certainty. It is that you can sit in front of a panel, describe what you searched, in which vocabularies, over what period, and explain why your claim still stands.
Which means keeping a record while you search. What you typed, where, and what came back. It takes minutes at the time and it is very difficult to reconstruct later. That record is also most of your literature review's method section, so you are not doing extra work, you are doing it early.
Where Gapnostic fits
This is the check Gapnostic was built to run. You describe your idea, it searches the published literature, and it reports how closely your idea overlaps with what already exists, what appears to be genuinely new, and which papers you need to read before you commit. Every judgement carries the paper it came from, so you can check the reasoning rather than trust it.
It does not decide for you, and it does not write anything. It does the searching and the cross-checking so that the decision in front of you is an informed one.