The term "validity" is thrown around a lot but there's no universal measure of psychometric validity. In simple terms, something is valid if it's predictive of something else that is useful (usually a behavior or another construct). There's plenty of studies that investigate the validity of love languages around things like marital satisfaction etc. More importantly, for the purposes of this study which is just based around self-reported preferences (and not psychometric properties), validity doesn't really matter. Reliability does, and there are is plenty of research around the reliability of these scales.