We know that latent variables represent theoretical concepts, but sometimes we need extra constructs representing other latent variables. For example, a questionnaire(For Example, General ability) may consist of several components(Verbal ability, Numerical ability, and Spatial ability), each of which components several questionnaire questions. In this case, the researcher can build the model based on the components of the questionnaire; in the term, it is called Higher-Order Construct Model or Hierarchical Model. In this case, there is a latent variable(For Example, General ability) in the model, which consists of all the questions in the questionnaire. Also, there are some other latent variables(Verbal ability, Numerical ability, and Spatial ability) that are made up of part of the questions in the questionnaire.
You can easily fit these types of models in SEM-PLS software.
Analysis By SEM-PLS 2022: https://statisme.com/Learn/PathAnalysis