No, it's not "talkings". Definitons of science are not based on replication/reproducibility.
Harder one, Popper's, says that a theory must provide a verifiable/falsifiable prediction, and it should be correct. (Problem is, there are too many probabilistic things, so you must lower the bar, by checking error rate/margin.) But not everything is verifiable or falsifiable strictly to this definition.
Then it's Kuhn's definition: a systematic observation and description of nature phoenomenons in own strict terms. Anthropology and sociology are exactly in this department.
Just an example what discovery sociologists made in mid-XX century: poor people tend to learn habits that lead them to risky and reckless behavior, stop them from getting education, etc. (What's the cause is another big talk, too broad for this thread.) And just talks by intellectuals would be like "they're just fools", "they have some genetic deficience", or maybe "they just haven't money, with enough money they'd be fine". That's talks.
Popular psychology that drops broad and seemingly profound judgements and commands, but that have nothing of substance -- that is talks too.