Cultural Similarity

[_John Larsen]

Cultures are, of course, very complicated things with many facets. Culture includes the collective experience of the people wrapped up in language, legend, fable, geography, government, literature, art and individual roles that the people employ. If it were possible to quantify all of the sub-elements of a culture, there would be literally millions upon millions of data points and propositions, both substantive and incidental.

Because of the large number of these cultural elements, it is highly probably that many of them will line up if you pick any two cultures at random and compare them. Statistically speaking, the huge number insures that there will be 100s of cultural elements that appear shared—but the appearance is deceiving since the two cultures arrived at the similar ideas from two completely different positions. Looking at these shared elements would give the distorted view that the cultures are specially related or derived one from another when that case is not justified.

For example, in linguistics there is a finite set of distinctive sounds the human being is capable of producing. Phonemes are produced by combining these sounds into distinctive elements of meaning that can be audibly differentiated. Because of the limited set, if you compare any two random languages, there will be 1000s and 1000s of phonemes that appear in both languages—the vast majority having no meaning in common. However, in any two languages there will be a handful of words where both the sound and the meaning correspond—at random. Because of the limited set of sounds and definitions, this is bound to happen and has been noted many times by linguists in many different languages.

Additionally, human experience is limited and all cultures face many of the same problems of living in this world. So an even higher degree of correlation would be expected than purely random, since our anatomy, biology, and even the weather will lead humans to make similar decisions. Likewise, a diffusion effect can occur among cultures that have no direct contact. The Spice Route comes to mind which extended from Rome into the Alaskan territory, although no one person was aware of the extent.

So if one wishes to show that a particular culture was spawned from or influenced by anther particular culture, more than the baseline expected amount, one must first show that the lineup of cultural and linguistic elements falls outside the normal distribution of randomly expected elements. In other words, if you took two complex cultural artifacts, and compared them, you might be able to find some surprising similarities. But unless you can demonstrate that the volume of those similarities is sufficient to overcome the null hypothesis of random or incidental cross over, you have not established causality.

[_maklelan]

So in response to Brian Stubb's article about over 2000 elements of Semitic and Afro-Asiatic languages found in Uto-Aztecan (which even takes into account systematic phoneme shifts like the unvoicing of voiced stops and the velarization of labial stops and glides in specific languages), what would you say?

PS - It wasn't me who noticed the phoneme shifts, it was the former president of the Linguistic Association of Canada and the United States. He says it cannot be attributed to coincidence.

[_John Larsen]

So in response to Brian Stubb's article about over 2000 elements of Semitic and Afro-Asiatic languages found in Uto-Aztecan (which even takes into account systematic phoneme shifts like the unvoicing of voiced stops and the velarization of labial stops and glides in specific languages), what would you say?

PS - It wasn't me who noticed the phoneme shifts, it was the former president of the Linguistic Association of Canada and the United States. He says it cannot be attributed to coincidence.

My whole point is that fact is unevaluateable unless you know what the base line expected amount. It might be that 2000 is high, low or in between. If you pick two languages at random, how many would you expect. I don't care what some former president says, he needs to show statistical significance.

[_maklelan]

So in response to Brian Stubb's article about over 2000 elements of Semitic and Afro-Asiatic languages found in Uto-Aztecan (which even takes into account systematic phoneme shifts like the unvoicing of voiced stops and the velarization of labial stops and glides in specific languages), what would you say?

PS - It wasn't me who noticed the phoneme shifts, it was the former president of the Linguistic Association of Canada and the United States. He says it cannot be attributed to coincidence.

My whole point is that fact is unevaluateable unless you know what the base line expected amount. It might be that 2000 is high, low or in between. If you pick two languages at random, how many would you expect. I don't care what some former president says, he needs to show statistical significance.

As a student of languages, specifically Semitic, I can tell you 2000 similarities, including borrowings (lexical, morphological, and phonemic) and shifts, is significant. I am not aware of any study which has been done specifically to address the question of the coincidence/relationship threshold, but I've certainly never seen similarities of this quality and degree occur coincidentally in two languages. I would invite you to present such a case, but I highly doubt you're any more privy to this information than I am.

While you can make the case that a relationship cannot be firmly concluded, you certainly can't argue that a relationship is at all precluded because of a lack of statistical significance, especially when that lack is due to the lack of research in that area. You will not find two unrelated languages with the same number and quality of similarities, though. That I can guarantee you.

[_John Larsen]

My whole point is that fact is unevaluateable unless you know what the base line expected amount. It might be that 2000 is high, low or in between. If you pick two languages at random, how many would you expect. I don't care what some former president says, he needs to show statistical significance.

As a student of languages, specifically Semitic, I can tell you 2000 similarities, including borrowings (lexical, morphological, and phonemic) and shifts, is significant. I am not aware of any study which has been done specifically to address the question of the coincidence/relationship threshold, but I've certainly never seen similarities of this quality and degree occur coincidentally in two languages. I would invite you to present such a case, but I highly doubt you're any more privy to this information than I am.

While you can make the case that a relationship cannot be firmly concluded, you certainly can't argue that a relationship is at all precluded because of a lack of statistical significance, especially when that lack is due to the lack of research in that area. You will not find two unrelated languages with the same number and quality of similarities, though. That I can guarantee you.

How can you tell me that is significant? Which two languages do not have 2000 similarities?

[_maklelan]

How can you tell me that is significant?

Because my specialization is comparative Semitics and because I speak four other languages. I've had plenty of experience comparing languages and taking note of similarities.

Which two languages do not have 2000 similarities?

I think telling you that you won't find any languages with the same quantity and quality of similarities is a clear indication of which two languages do not have 2000 such similarities.

[_CaliforniaKid]

Hi John,

I think you're on to something here. A while back at MADB we discussed ways of determining the statistical significance of Book of Mormon names with authentic Hebrew parallels. Until such a study is performed (and it would be quite the undertaking!), we cannot really be sure whether the parallels that have been thus far identified are actually significant.

Best,

-Chris

[_harmony]

I feel like I dropped into the middle of a conversation and have no clue what the preceding points were.

What are you talking about? And where can I find the thread that explains what you've already discussed?

[_CaliforniaKid]

Here is the thread I mentioned in which we discussed methodologies for determining the statistical significance of Hebrew parallels to Book of Mormon names.

[_John Larsen]

How can you tell me that is significant?

Because my specialization is comparative Semitics and because I speak four other languages. I've had plenty of experience comparing languages and taking note of similarities.

Which two languages do not have 2000 similarities?

I think telling you that you won't find any languages with the same quantity and quality of similarities is a clear indication of which two languages do not have 2000 such similarities.

You think so eh? Unless they have changed the way the do linguistics since I graduated, your opinion doesn't matter for much. Nothing personal. Science is based on evidence and method. Not on well educated hunches--unless those hunches are validated through experimentation.

Once again, how do you know that your "2000" similarities are not then norm. So far the only answer you have given is that you happen to "speak" 4 languages.