Dear Sir,
While essentially agreeing with your views, we would like to put the language differently, as your language can be misleading. In our essay this year, we have defined language as the transposition of information to another system's CPU or mind by signals or sounds using energy (self communication is perception). The transposition may relate to a fixed object/information. It can be used in different domains and different contexts or require modifications in prescribed manner depending upon the context. Since mathematics follows these rules, it is also a language, but it talks of only quantitative aspect of Nature. You are welcome to read it and comment.
In our essay: "INFORMATION HIDES IN THE GLARE OF REALITY http://fqxi.org/community/forum/topic/1776" published on May 31, 2013, we had discussed this subject elaborately. Perception is the processing of the result of measurements of different but related fields of something with some stored data to convey a combined form "it is like that", where "it" refers to an object (constituted of bits) and "that" refers to a concept signified by the object (self-contained representation). Measurement returns restricted information related to only one field at a time. To understand all aspects, we have to take multiple readings of all aspects. Hence in addition to encryption (language phrased in terms of algorithms executed on certain computing machines - sequence of symbols), compression (quantification and reduction of complexity - grammar) and data transmission (sound, signals), there is a necessity of mixing information (mass of text, volume of intermediate data, time over which such process will be executed) related to different aspects (readings generated from different fields), with a common code (data structure - strings) to bring it to a format "it is like that".
Obviously, as you have said, there is information defining the object outside the object itself. That is information, but it is related to the object through similarity. The description of your example is not appropriate. While analyzing the speed of a car, we do not compare the object 'car' with the object 'road' and exclude everything that is around them. We measure the interval between two events (position of car on different points on the road). This interval is compared with a standard interval (standard unit) and generalized over the total duration to get speed of the car. Everything is interdependent and interconnected with everything else in one way or the other. Thus, as you say, dissociating something from others "can be difficult to make or impossible". Whenever we do that, we must be "losing information in the process". But then information must be real, which corresponds to provable options that are knowable and describable using language. A computer can only compare, but cannot "know". If can obey commands to transmit information, but it cannot utilize language to transpose information to another system's CPU or mind.
Regards,
basudeba