“Definitely” is an interesting word. We often use it casually to just indicate strong ascent or certainty - the absence of doubt. Or it can also act as a kind of synonym for “unambiguously” - that something is clear and precise.
Justification in epistemology is an interesting beast, since it often seems right to say that one can be justified in having a belief without that belief actually being veridical. You were perfectly reasonable in drawing the conclusion that you did that there was a hay barn some distance ahead of you, because you weren’t to know that the farmers around here like to make big wooden stands and paint them to display barn facades to keep the tax men on their toes. Why would you think that in a vacuum? Nobody thinks anything less of you because you were deceived into thinking something about the presence of a barn - the whole purpose of the facade is to do just that.
But, here’s the thing. The evidence that led to your belief clearly wasn’t unambiguous, regardless of whatever strength of belief you might have in your belief. The visual information you received was equally consistent with the thesis that it was a finely crafted illusion. It’s just that, most of the time, this possibility is totally irrelevant to the situation, so we (generally speaking) learn to discard that possibility. So having a justified belief here isn’t the same as a “Definite” belief; there were some alternative possibilities that you just decided were less likely.
Importantly, adding the truth condition does not change this. To see this, consider - what if there really is a hay barn there, but it’s completely hidden from sight by the facade? This would make your belief both true and justified, and yet there seems to be a kind of lucky coincidence at work that made those two line up; you never actually saw the real barn, so how could you be said to know that the barn was there on the grounds of your visual evidence?
This is a thought experiment popularised by Alvin Goldman in his paper “Discrimination and Perceptual Knowledge”. Goldman was a proponent of a Causal theory of knowledge, which (in short) argues that it’s not just that the belief in question corresponds to a fact (I.e. is True) but also that it has to have been in some way causally connected to that fact - that your “reason for believing” is in fact tied to the actual state of affairs that makes your belief true.
This view puts a lot of weight on the concept of causation, so in classic terms we might call this a “rationalist” view of scientific knowledge. For a more “empiricist” take on the same underlying idea, we could instead suggest that the kind of justification involved in scientific knowledge attributions requires a greater sensitivity to the context of utterance.
Basically, as Fred Dretske emphasised, we also need to ask ourselves about those alternatives that might be relevant to the situation we find ourselves in. Our justification might not be definite - the question is, are the ways in which our justification has possible holes salient, and if so, our scientific analysis must account for them. We might be perfectly justified in thinking there’s a real hay barn there, but if we know we’re in “fake barn county” then a suitable amount of skepticism ought to be brought to bear, in addition to what might count as evidence the rest of the time.
The practice of Science is a context of particular scrutiny, and so a great many alternatives could well be relevant to any given thesis. A scientist doesn’t just seek good reason to believe themselves - they also seek models making sound predictions, productive explanations and regarded publications! So on this view, it makes sense that Scientific knowledge ought to be taken as more rigorous than knowledge of other forms.
There’s plenty more to explore on alternatives to the Tripartite model; hopefully what my attempt at an intro demonstrates is the space into which one person’s “definite” beliefs might well be too presumptuously strong for a theory of scientific knowledge.
