Reviewing a set of high-
[I gratefully acknowledge the assistance of NASA and Jet Propulsion Laboratory in providing high-
My original “text” was quite short, occupying only 0.02 square centimeters (which may account for the fact that NASA researchers overlooked the linguistic evidence that I found). Nonetheless, considerable evidence of linguistic patterns became apparent to me after close investigation and comparison to patterns in adjacent portions of the rock surface. In Example (a), I quote the first “text” in its entirety:
955 967 989 919 912 913 916 998 966 988 956 978
where each three-
Through comparative analysis of thousands of other “texts” in a surrounding region of 4.5 square centimeters, I arrived at the following English gloss for Example (a) above:
May the mountain be yellow.
Comparative analysis of subsequent texts yielded dialogues such as Example (b).
Speaker A: No, I did not go to the valley. Speaker B: No... [unintelligible] Speaker A: I am of the yellow fields of the valley. Speaker B: The fields... The yellow fields... What is yellow? Speaker A: Yellow is not red, not blue, not green. A yellow star is yellow, is it not? Speaker B: Yellow is yellow. Speaker A: I am of the yellow fields of the valley. Speaker B: Yes, my friend, of the yellow fields of the valley we are.
“Speaker A” and “Speaker B” are considered different speakers since the utterances form clear patterns of dialogue. Note: The present study focuses on lexical/discourse effects, so morpheme-
In a three year study, I have been able to analyze 105,512 lines of discourse data based on 4.5 square centimeters of the rock surface. Using comparative techniques across these thousands of lines of text, I have glossed the meaning of each line and have begun developing the analysis outlined below.
First of all, the language is based entirely on color wavelengths. I have identified contrastive color wavelengths for every morpheme in the data. However, the language also allows for a surprising level of variability; meaning is continually negotiated on a morpheme-
Speaker A: And then I climbed the mountain. Speaker B: You climbed the what? Speaker A: Mountain. Large. Tall. The large thing between two valleys. Hides the stars. A mountain is a mountain, is it not? Speaker B: A mountain is a mountain. [accepts new wavelength for mountain] Speaker A: And then later I saw a dusty trail. Speaker B: You saw a what trail? Speaker A: Dusty, tiny specks of dirt. Dry. Not mud. Dusty is dusty, is it not? Speaker B: Dusty is dusty. [accepts new wavelength for dusty]
In the lexical negotiations in Example (c), Speaker A’s word, e.g., dusty, has a slightly different color wavelength than other usages of dusty in prior texts, generally differing by about 10-20 nanometers. In each case, Speaker B accepts the new wavelength for the lexeme only after the definition has been explained by Speaker A (such as, “Mountain. Large. Tall. Hides the stars. A mountain is a mountain, is it not?”). Once this explanation has been accepted, Speaker B then continues to use this wavelength for mountain. However, only 12 lines later into the dialogue, the lexeme changes again.
Speaker B: I long for the mountain of the East Speaker A: You long for what of the East? Speaker B: Mountain. Tall. Large. One climbs it. A mountain is a mountain, is it not? Speaker A: A mountain is a mountain. [accepts new wavelength for mountain]
Yet just 7 lines later, Speaker A changes the lexeme for mountain once again:
Speaker B: Would that the rain stop soon. Speaker A: The old ones on the mountain have said so. Speaker B: The old ones on the what? Speaker A: The old ones on the mountain. Mountain is mountain: Large, not a valley, towering in the sky, difficult to climb, hides the stars... Mountain is mountain, is it not? Speaker B: Mountain is mountain. Speaker A: The old ones on the mountain have prophesied that the rain will stop. Speaker B: Yes, that the rain will stop.
Such lexical negotiation is not limited to nouns and adjectives. Example (f) shows variation in pronouns as well:
Speaker A: And I am told that it was on a market day that you fled to the town on the mountain. Speaker B: Who fled to the town on the mountain? Speaker A: You, not they, not I, not s/he, but you — the one with whom I speak. You are you, are you not? Speaker B: You are you. I am I. Speaker A: It was on a market day that you fled to the town on the mountain. Speaker B: Yes, that is the day that I fled.
Note, however, that even in the span of this brief dialogue, the pronunciation (i.e., color wavelength) of mountain changes yet again: The dialogue in Example (g) immediately follows Example (f).
Speaker B: A market day. Yes, I began at the river and fled to the East, all the way to the Great Mountain. Speaker A: You fled to the Great what? Speaker B: The Great Mountain. Mountain is mountain. Tall, not a valley. You have spoken of it before. Hides the stars. Difficult to ascend. Speaker A: Mountain is mountain. Speaker B: I fled to the Great Mountain. Speaker A: ...[unintelligible] the mountain of greatness, of all the people, of our safety. Speaker B: Of whose safety? Speaker A: Of our safety. Our is our. You and I together. Our house, our family, our home. Our is our, is it not? Speaker B: Our is our. It is the mountain of our people, of our safety. Speaker A: What is ‘it’? Speaker B: It is it. Not you, I, they. Not s/he, but it. It is it, is it not? Speaker A: It is it. Yes, the... [unintelligible] And now the rain comes. Speaker B: The rain comes. Let the rain come, and do not fear. Speaker A: What is ‘not’? Speaker B: Not. The negative, untrue, unreal, in the mind rather than reality. Not. Not is not. Speaker A: Not is not. Yes, not is not. I do not fear the rain.
Such lexical negotiations permeate the entire 105,512 lines of text in the corpus. In fact, lexical “explanations” of this type occur in an average of every 6 lines between these two speakers, and occupy a full 34% of all speech in this data set.
My initial assumption was that the two speakers were using different dialects and therefore were frequently encountering simple communication problems due to lexical mismatches between two dialects. However, this analysis is not supported by the discourse data. In particular, I note that speakers often use the same wavelengths for the same words moments before in the text, and yet the same speakers find that they must explain the meaning once again in order for the dialogue to continue. For example, Speaker A uses 977 nm for right at one point in the text, but the word right then changes to 988 nm about 15 lines later, where the two speakers must agree once again on a new pronunciation. 12 lines later, Speaker B shifts back to 977 nm and Speaker A asks for clarification. Thus, even though Speaker A uses the same wavelength for right just 27 lines earlier, s/he still must ask for clarification again since the wavelength has subsequently changed.
Therefore, I conclude that synchronic dialect variation is not an adequate analysis. Rather, I suggest that this is a case of extremely rapid diachronic change (ERDC). In this ERDC language, the change is “extremely rapid” in the sense that lexical definitions change within a few lines of interpersonal dialogue. This suggests a contrast with the commonly held view of language change as a long-
Tim Pulju (p.c.) rightly points out that such a language is not an efficient mode of communication. One normally expects natural languages to develop in such a way that straightforward communication is at least generally smooth and relatively efficient. On the other hand, it may be that the speakers in this Mars Lander corpus have (a) highly malleable mental structures that do not allow for fixed lexemes (i.e., unlike human physiology, which supports considerably more stable cognitive structures and a sense of a “fixed” lexicon, albeit slowly changing over a span of years), or perhaps (b) the chemical processes required in this color-
Further study may resolve such questions. In the meantime, I suggest that this language (which I have called “955” due to the wavelength of the first morpheme in the corpus) is an example of an ERDC language, and I would anticipate the discovery of many more such ERDC languages in the future.
Finally, given this result, it is hoped that this study will encourage linguists to move beyond the narrow, Earth-