From the introduction, "Our belief that translation is possible, or, for that matter, that human beings can communicate at all, is based upon two presuppositions: 1) that human beings everywhere live in approximately the same world of experience, and 2) that all human beings have approximately the same mental and psychological apparatus.
But we must at the same time recognize that all human beings must abstract to a high degree from the complex environment around them which constantly bombards them with great amounts of sense stimuli. We are required to impose order on, or find order in (depending upon your point of view), the world which we perceive through these sense stimuli. It is also clear that people of different cultures, or for that matter individuals within the same culture, abstract in noticeably different ways. Take for example the way in which people around the world vary in terms of their degree of differentiation of "things" like color, snow, horses, plants, etc. Notice also the different patternings that people perceive in the same perceptions, as for instance in the well-known Rorschach tests. There is also some evidence that languages differ in the predications which can be made on the same observable events. For example, the verb root in Copala Trique which is used to describe the same observable phenomenon which is described in English by the word "cover" as in "The woman covered the baby with the blanket", actually predicates a different abstracted event than does the English root. In Trique, the corresponding sentence, in glosses, reads "The woman covered the blanket to the baby." That is, whereas in English the event of something being done to the baby, with the blanket as instrument, is predicated, in Trique an event of something happening to the blanket (i.e. being moved to the top of the baby) with baby serving as the new location of the blanket, is predicated. From this we must conclude that we cannot know what the "world of experience" really is, especially if we hope to go about our search merely by asking individuals what it is that they perceive. But, in linguistics we do not claim to be dealing with the "real world" but only with the meaning areas and patterns which people abstract from the world of experience and with the manner in which they convert these meaning patterns into a linear phonetic output. And, as a matter of fact, we claim that, although it is not possible to get inside of a man's mind to see what processes go on there, we can get some idea of what the bits of meaning and relations are which people actually do seem to employ and which are encodable into sound. We do this by the scientific processes of building explanatory models, drawing upon the resources of introspection and observation in others of language behavior. These models posit what we understand to be the raw (semantic) material behind the phonetic output we can most directly observe and what processes are employed to convert that raw material into the observed output."
Hollenbach, Bruce E.
"Categories of verbal ideas and case relations,"
Work Papers of the Summer Institute of Linguistics, University of North Dakota Session: Vol. 17, Article 13.
Available at: https://commons.und.edu/sil-work-papers/vol17/iss1/13