Differences in relationships between concepts
(a) Generic-specific: the relationship of one concept as being more generic and another as more specific is reflected in the lexical structure of all languages and their taxonomies (e.g. in the case of plants or animals). There are situations in many languages where the same word form is used at several different levels within such taxonomies (e.g. man). The generic-specific relationships of two languages tend to be quite different. For example, in the Philippines "rice" is the generic term for all forms of grain so that "wheat" might be translated as "rice called wheat". Slavic languages do not have separate words for arm and hand which are together denoted by the same term. In Isnag (Philippines) the "trunk" of a tree is thought of as being in two parts separately named. The Tlingit of Alaska have no general word for "swim", instead they have many specific words for different kinds of swimming. Larson points out that languages tend to differ most in generic terminology, rather than specific. Whereas equivalent words can be found for specific objects or for phenomena such as "murder", "lie" and "steal", it may be very hard to find equivalent generic words for "bad". The translation of such abstract terms is often very complex, especially if the cultural contexts of the two languages are quite different. It is to be expected that the complexity increases with the degree of abstraction. 
(b) Synonyms-antonyms: A second language may not have a specific word equivalent for each of the synonyms of the source language. There may be more synonyms or less. All languages have pairs of words which are antonyms, but different languages have different sets of antonyms. Larson gives the example of the two antonyms short/tall (vertical) and short/long (horizontal) which are covered by one Aguarana antonym set. Some languages have a word for only one of a pair, the other being indicated by the negative. 
(c) Contrastive pairs: In all languages there are pairs of words which differ from each other only by a single component of meaning. Larson gives as example "show" and "see" in which show has the additional meaning of "cause to" (see), similarly in the case of "drop" and "fall" or "make" and "be". It is not uncommon that a language will have no exact equivalent for "show", "drop" and "make", but some causative form will be used instead. Two languages may have the same concept set as far as the generic components distinguishing each word from others in the set will be different. There may be more lexical items in the set and the contrastive components may not match. Larson gives the example of the set "human" which in English has components man, woman, boy and girl. In Aguarana the "man" term of this set must be distinguished as either married or unmarried.
(d) Semantic sets: The lexical items of a language represent a network of interrelated meanings that has been called a cognitive network. No two languages will have equivalent sets of terms referring to a particular domain. This is clearer in the case of tangible objects, but is also true in the case of verbs. Larson gives the example of Bora (Peru) in which a number of verb roots are used to describe different forms of "coming" and "going". Which do not match the English verbs. The roots include: go, go to, going arrive at, come, come to, coming arrive at, and come back to.