A very attractive feature of propositions, if defined as the meanings of sentences, or in some such way, is that they are language-independent. Thus the sentence water is transparent and the sentences in all other languages with exactly the same meaning correspond to, or 'are', one and the same proposition: the proposition water is transparent. Altho every language may have a sentence with a meaning identical to this proposition water is transparent, it is not hard to conceive of some sentence in this language for which there is no exact equivalent in a particular other language (not belonging to the same cluster of 'languages'), and vice versa. In that case there is a proposition which can be uttered in this language but not in some other language, and a proposition which can be uttered in some other language but not in the language which is our present means of communication. The next step is then to conceive of a proposition which cannot be uttered without change of meaning in any language. Thinking (not speaking or writing) along the lines of such propositions (if possible) would be as language-independent as propositions are themselves: words may be the instruments of thought, that does not imply that thought could not do without words.

The question whether thoughts can be entirely language-independent is closely related to the question whether the communication of thoughts, and of other elements of propositional attitudes, such as feelings, can be entirely independent of language. As the transmission or exchange of thoughts and feelings can take place outside the verbal medium of a spoken or written language, it seems evident that propositional communication is at least not wholly language-dependent. Yet, our present means of communication, and also the most important or easy means of communication (in the present context), is verbal communication in a particular language. It is therefore of great import to have some idea if, and to what extent, the expression of thoughts and feelings depends on the kind of linguistic medium used, given that such a medium is used. After all, we may prefer to plane away the rough and defective parts of the language employed, rather than to alter our thoughts themselves because we could not express them properly and easily by means of a tool still largely controlled by the invisible, dead hand of tradition.

That there is a relationship between language on the one hand and people's thought and behavior on the other, is generally acknowledged. Thus it has been stated that 'we think the way we speak', and it has been added to that that 'we speak the way we think'. A special aspect of this 'speaking likeness' is the connection between language and 'ideology' in the broadest sense of the word, including both political and nonpolitical, religious and nonreligious ideologies. If the interplay between language and thought is not entirely illusory, a dissimilar ideology must in some respect correspond to a dissimilar linguistic usage (maybe just a different connotation for some words), especially in a field it is primarily concerned with. For example, it has been argued that one and the same labor situation may be described in completely different ways dependent on one's political or economic point of view. The active person giving or doing something in such a situation may be the 'worker' for the one and the 'employer', or the person said to give work, for the other, whereas the passive person receiving something is the 'capitalist' for the one and the 'employee', or the person said to take work, for the other. (The difference in usage is even more clear in languages in which the employee is labeled "work-taker" and the employer is termed "work-giver".) Now, the truth may be that in a particular labor situation the same person gives one thing (for example, work) and takes another thing (for example, labor power) but the exclusive emphasis on one party's giving and the other party's taking is ideological. The difference in terminology, that is, language, therefore bespeaks a difference in perception, and with it a difference in thought, even tho the conditions are the same.

It is sometimes said that language is not only a means of communication to tell other people something about reality, but also a means to create reality, or rather one's perception of reality. Words can make and break such perception. Tho language is itself a product of society, it has been pointed out that it can alter the relations between people themselves as long as the right linguistic tools are used. As the prime instrument of communication language is also the vehicle for the spread of ideologies (and as a general means this applies to both good and bad ideologies). But as a product of society language is itself already infused with the spirit of past and present thought which may be compatible or incompatible with a new or future doctrine. And altho language may not create ideology since ideology only manifests itself in language --as has been argued-- this does not refute the argument that language perpetuates a particular ideology or way of thinking. When a present-day language still draws traditional distinctions which are the remnants of exploded ideas, or when it draws no relevant distinction because it never did so before, it goes on creating an impression of normality and abnormality which is counternormative. The sanctioning of such an impression is, then, political, religious or otherwise ideological. Language is just too often lagging behind new developments: the social relations (as between the sexes) may have changed while linguistic usage has not yet been adapted. That is why some characteristics of a particular language may be relics which do not so much reflect a present, but a former world-view of the community of speakers of that language.

Research has shown that words determine to a certain extent the organization of what is perceived, that they 'stabilize and confirm such an organization'. And conversely, when certain differentiations have become or are deemed important, these differentiations make their way into the verbal categories of the language. The most obvious reason for a difference in such categories is, of course, a difference in the conditions different linguistic communities are living under. Yet, some people do not have 'sibs' (only sisters and or or brothers), and other people do not have 'parents' (only a mother and a father); nevertheless one would say that (at least in this respect) the conditions are the same in the linguistic communities concerned. It has been said, however, that it is not inherent in a particular language to provide us with a simple and short verbalization. Having to say "one or more sisters and or or brothers" would, then, only be a question of having to describe one's conditions in a more or less cumbersome way, and the (needed) length of verbalization would depend in principle on the context. For example, we may describe a figure just as "a rectangle" (without mentioning its position, size, and so on), and we may describe a color just as "green" (and not as "a little bit yellowish green"). But these two examples differ essentially from each other, and context-dependency is therefore too vague a concept to explain or justify the length of verbalization with.

Being-a-parallelogram, having-a-certain-position and being-a-certain-color all represent a factor which --let us say-- admits of degrees. (For the sake of convenience we assume that the factors are one-dimensional.) Now, one can determine an angle of a parallelogram as accurately as one wishes, and one can locate a color in a spectrum more or less accurately (altho the meaning of color terms always remains arbitrary to some extent) but in each case one has supposed that the factor being a parallelogram and the factor color are relevant. The accuracy of the description of this one quality is then further dependent on the same contextual features which make the factor itself relevant. The question whether the introduction of position, size, and other factors (not just degrees of accuracy) into the description is required, may depend on entirely different contextual features tho. If they are, one adds an affix, an adjective or adverb, or an adjectival or adverbial, subordinate clause (in the present language and in languages with a similar structure).

Changing a description from sib to sister or brother, or from parent to mother or father, introduces the factor gender. Granted that everyday language is not obliged to produce short verbalizations everywhere, and that it can only have a limited vocabulary, a systematic procedure would be to denote a sister with female sib and a brother with male sib in those contexts in which gender matters. But there are but too many languages where it is the other way around: they have words for brothers and sisters but not for sibs. Or, there is a word for sibs, but it is rarely employed in comparison to the words brother and sister. While it is a truism that language is cumbersome at one place and efficient at another, it may be quite interesting to find out where it is cumbersome and where it is short and clear. If it is not short and clear until a new factor has been introduced, then it carries the inherent presupposition that that factor is relevant in every context. This implies that language can have the tendency to make appear relevant, and to make people believe that something is relevant that is actually irrelevant (except under specific circumstances). And this is much more serious than the converse, because it underlies stereotypical thinking and the disbelief in equality (as we will see later). (Altho in this regard the sister/brother dichotomy may be innocent compared with the asymmetrical girl/boy, spinster/bachelor, queen/king and similar traditional juxtapositions.)

People's acting and thinking is responsible for the emergence of language, but once it is there, it is this very language which, in turn, can have a great impact upon people's thinking and acting. Obviously it depends on what language what that impact is, and how strong or far-reaching it is. Like (almost?) all instruments also language can be used for praiseworthy and for blameworthy causes. The point is to nail its implicit ideological prejudices and presuppositions, and to counterbalance it where it is askew. It has already been said before that change of language, or rather an improvement of language, is certainly not unimportant for anyone who has to transmit new knowledge in the field of training, education, cultivation, and the like. The right use of language will promote that people actually discern what is there, and that they will act accordingly. This may either involve a difference where no distinction was drawn before, or no difference where an irrelevant or false distinction was drawn before.

As has been argued correctly, people are not always and not solely objects of their language. Those who create language themselves may not change the ground-facts, yet they do somehow 'participate in the formation of the world', even if only in the creation of new facts of thought and correspondence. And creation often involves destruction: fettered by traditional thought, these language users must employ the weapons of thought to destroy the fetters of tradition.