It is usual to consider that standards generate mixed feelings among scientists. They are often seen as not really reflecting the state of the art in a given domain and a hindrance to scientific creativity. Still, scientists should theoretically be at the best place to bring their expertise into standard developments, being even more neutral on issues that may typically be related to competing industrial interests. Even if it could be thought of as even more complex to think about developping standards in the humanities, we will show how this can be made feasible through the experience gained both within the Text Encoding Initiative consortium and the International Organisation for Standardisation. By taking the specific case of lexical resources, we will try to show how this brings about new ideas for designing future research infrastrcutures in the human and social sciences.