The Menzerath-Altmann law can be traced back to 1928, when it was observed that long German words tend to consist of short syllables, and vice versa. The law was later generalized to be applicable to all languages. In essence, the Menzerath-Altmann law constitutes a mathematical model for the relation between sizes of linguistic constituents and constructs. It is almost ubiquitous, both across languages and across different types of linguistic units, and can be expressed by a simple mathematical formula. The book presents not only a summary of the results achieved so far but also new findings. It identifies some of the conditions under which the law is valid, offers an interpretation of the model parameters, and extends the scope of the law across the hierarchy of linguistic units. It contributes to the development of theoretical linguistics and to mathematical modelling of language properties.