Thursday, March 29, 2007
More Random Copying
The press is starting to pick up the Random Copying story. To the right is a graph of the rate of occurrences over time since the story first "broke" yesterday at about noon. Let's see if the popularity of the story takes the shape of a random variant. It would be great to be able to do this on the fly - generating histograms of events from news.google.com based on queries. Google "trends" does this to some degree, but it uses search terms to build its data - not frequency of reporting in the news media. Thus its measuring something different - the rate at which people go to the web to look for a term -- not the rate at which it appears on the web. Building an application to do this shouldn't too hard given the automated way in which news.google.com search are generated. One would just need to parse for time and title of each entry (assuming each is really related to the topic of interest). Hmmmm.. I'll have to cruft something up. Maybe later.
Wednesday, March 28, 2007
The media is starting to pick up on the Bentley et al (2007) E&HB paper on random copying. Mostly it has been press in the UK (and India) but that could change tomorrow. It is kind of ironic that the message of the paper we wrote is now experiencing the "random copy" effect we describe as a function of the media passing the story on to one another. As it increases in popularity in the press, at least for the period of time before another story pops, it will continue to increase in frequency on science news sites. The frequency of the story itself is predicted by the story. Funny. Anyways, these are the media references to date. http://www.inthenews.co.uk/infocus/features/in-focus/changing-fashion-fashionable-$1072208.htm http://www.spiritindia.com/health-care-news-articles-7830.html http://www.livescience.com/humanbiology/070328_fickle_fashion.html http://newswire.ascribe.org/cgi-bin/behold.pl?ascribeid=20070328.094246&time=11%2006%20PDT&year=2007&public=0 http://news.sky.com/skynews/article/0,,30000-1258130,00.html
Monday, March 26, 2007
One implication of the random-copying model paper is that many of the cultural phenomena that we see are actually a reflection of the effect of random copying between individuals. We like to think that we strategies what we do or that we have a good "reasons" for what we do. While people may have all kinds of rationale for what they do, the data consistently show that change is a function of copying. To the right is another example of this effect. The graph shows the incidence of reports of "UFO" sightings per day from June 1 through July 30, 1947. The distribution is consistent with a trait moving through a population copying in a random fashion. I suspect the "witch sightings" data from the Salem during the late 17th century take the same distribution. Could incidence of mentions of WMD in Iraq by the Bush administration be the same?
Evolution and Human Behavior
It looks like the Bentley, Lipo, Herzog and Hahn paper Regular rates of popular culture change reflect random copying paper is going to published in Evolution and Human Behavior some time this week. You can download a copy of the paper here: http://www.csulb.edu/~clipo/papers/BentleyEtAl-2007-RandomCopying.pdf We are also doing a press release for this through CSULB and the University of Durham. Here's the text of the PR that is slated to go out shortly. (Cal State Long Beach draft v. 3/26/07) March 27, 2007 #2007-XXX Random Copying Influences Popular Trends More Than Rational Choices, Says International Study Group Including Cal State Long Beach Professor Identifying the next trend-forward thing can be a crucial prediction worth billions of dollars to marketers. Yet, the popularity of things like baby names, music, dog breeds or fashions will change at a constant rate regardless of population size through a process of people randomly copying trends, according to a new study by an international team of academics including a California State University, Long Beach professor. Controversially, this contradicts classic economic models which believe that people make rational choices about the clothes they wear, the way they dance or the music they listen to. Researchers at Durham University, England, as well as CSULB, Western Carolina University and Indiana University authored an article titled “Regular rates of popular culture change reflect random copying” that appears in the May issue of the journal Evolution and Human Behavior. It shows that almost all of us are copycats and it is guaranteed that our taste in music or baby names will change at a consistent rate over time, but there is no way of predicting how it will change as it is completely random. New ideas become highly popular by chance alone, and then over time are replaced by others, all through the process of copying with occasional innovation. Carl Lipo, an associate professor in anthropology at Cal State Long Beach and research scientist at CSULB’s Institute for Integrated Research on Materials, Environments and Society (IIRMES), said that the results demonstrate how relatively simple models can be used to explain remarkably complex phenomena. “What we demonstrate,” he stated, “is that the aggregate effect of simple rules often underlie what we see as organization at higher scales. We don’t need to invoke the idea of some individual or group for patterns to emerge.” Led by Alex Bentley of the Durham University Anthropology Department, the team looked at the Billboard Top 200 chart and found that it turned over at a constant average rate for over 30 years between 1950 and 1980. The number of albums entering and exiting the chart varied from day to day and month to month, but overall the average stayed fixed at 5.6 percent per month for the full 30-year period. A similar consistent turnover rate was established for the top baby names and dog breeds. Bentley also is associated with the Centre for the Evolution of Cultural Diversity, funded by UK’s Arts and Humanities Research Council. Psychology Professor Harold A. Herzog of Western Carolina University and biology Professor Matthew W. Hahn of Indiana University also participated in the study. Their real-world data was matched by computer simulations of a random copying model with 2,000 individuals copying each other from one instant to the next, with a small proportion of innovators (two percent or less). During the simulation, they kept track of the Top 40, Top 100 and other lists of popular trends and monitored how much turnover there was. The model predicted continuous and regular turnover matching the proof of the real-world data from the charts of baby names, music and dog breeds. How quickly a list will change depends on the size of the list—the more choices people have, the quicker trends become popular or unpopular. However, the research has found that the size of the population does not have an impact on the turnover of lists. Although a higher population means more new ideas are out there, the turnover on a top 100 list does not increase as there is more competition for any particular idea to reach the list. Marketing professionals who use viral marketing, which spreads information by word-of-mouth through social networks and the Web, often classify people as innovators, early adopters and copiers. “Innovators are the cool ones who don’t bother imitating other people, but instead ‘pump’ new fashions into our world,” Bentley said. “Most are ignored, but some get copied. If the innovator is already a ‘cool’ celebrity, it means something shoots up in popularity much faster than you would predict likely via random copying. However, turnover over time will still be constant. “The model we have discovered predicts that the turnover of fashion will be proportional to the square root of the proportion of innovators, regardless of population size,” he noted. Since it is a game of chance, the model cannot predict how any one particular fad will fare, or which trends will become fashionable, just that new trends will definitely emerge at a regular and predictable rate. The discovery that change is continual and regular under the random copying model means it could be a useful tool to predict change rates as well as distinguish copying from other forms of collective behavior. There are areas in society where random copying is desired. For example, community campaigns to recycle waste benefit by people randomly copying each other’s behavior. However, in other areas such as politics, rational, informed choices are desirable. # # # For further information, contact:
- Dr. Carl P. Lipo, California State University, Long Beach, Department of Anthropology; Tel: +01 (562) 985-2393; e-mail: firstname.lastname@example.org
- Anne Ambrose, Public Affairs Office, California State University Long Beach, Tel: +01 (562) 985-2582, email@example.com; or Rick Gloady, Tel: +01 (562) 985-5454; e-mail: firstname.lastname@example.org
- Dr Alex Bentley, Durham University, Anthropology Department; Tel: +44 (0)191 334 6198;
- e-mail email@example.com
- Media and Public Affairs Office, Durham University; Tel: +44 (0)191 334 6075; e-mail firstname.lastname@example.org
There is a short article about my field work and research on Easter Island in the most recent issue of Smithsonian magazine. http://www.smithsonianmagazine.com/issues/2007/april/easter.php The comments by Joanne Van Tilburg are perhaps the most interesting as they demonstrate her tightly-held beliefs about the prehistory of the island and her reluctance to consider the evidence at it exists. She comes across as though she is putting the story before information, picking and choosing the "facts" as they fit her notions about what "must have" happened. For example, why is it inconceivable that people started making platforms when they first arrived on the island? Sure it makes little sense in 19th century "cultural evolution" framework in which all cultural change goes from "simple" to "complex" in a linear trajectory. But if groups were making platforms and other kinds of ceremonial architecture before they arrived, why would they necessary stop? In fact, this may be one of the first things they would do if the "function" of platforms is religious/honor of ancestor or whatever. The point is that we have to discard the badly constructed, largely data-free story that has grown up about the island (starting with missionaries) and generate information about the structure and distribution of the archaeological record. Until we do this task, the island will remain a "mystery" -- but perhaps that is ultimate goal of Van Tilburg in the end.
Friday, March 23, 2007
For all of you systematics in prehistory fans, philosophers of archaeological science, and mullers of conceptual frameworks for generating knowledge in the past, I just posted a couple of new RCD lectures. http://darwin.anth.csulb.edu/rcd/497/Lectures.html Enjoy.
Just for kicks, I took an old webcam I had and created a live web feed for my office. Why? I dunno. But you can now peer into the dull activities of an archaeology professor, 7x24. Whoohooo! http://www.csulb.edu/~clipo/webcam/
Society for California Archaeology Meetings - 2007
I am pleased to report that a number of students from the CSULB archaeology program are presenting papers at this year's Society for California Archaeology meetings in San Jose. The presence of students who present papers and posters at meetings is an excellent assessment measure of success of a program. I think we do remarkably well in this area. This year's SCA meeting participants are: Kristin Safi Luminescence Dating of Ceramic Samples from the Southern California Desert (Symposium: Current Applications of Advanced Technology in California Archaeology
- Luminescence dating can provide chronometric information about archaeological materials in regions where few cultural remains are readily available for the determination of temporal provenience. The desert region of southern California is one particular area where luminescence dating has great potential for providing chronometric information for aggregate scale archaeological features. In this study, coarse grain and fine grain analysis is conducted utilizing Optically Stimulated Luminescence/Single Aliquot Regeneration (OSL/SAR) dating techniques to determine the direct age dates on five ceramic sherds from two survey areas in Indio, California. Compositional analyses are also conducted to assess and calibrate the chronometric data to achieve greatinformation useful in provenance determination.
- Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) shows considerable promise as a minimally invasive technique for chemical characterization of rock art pigments. LA-ICP-MS offers high sensitivity to trace elements, requires only a small amount of sample with minimal sample preparation, and allows pigment to be differentiated from host rock to a degree not possible with other characterization methods. Analyses of paint samples taken directly from pictograph panels at Little Lake and of raw pigments previously recovered from excavations at the Stahl Site demonstrate the utility of LA-ICP-MS for making intra- and inter-site comparisons of pigment compositions.
Thursday, March 22, 2007
New RCD Lectures
Finally, I have freed up enough time to get more of the Dunnell 497 videos converted to DVD and to web-friendly movies. I have now posted Lectures 3 and 4 on my website at: http://darwin.anth.csulb.edu/rcd/497/Lectures.html
Sunday, March 11, 2007
One of the topics I talk about in my Archaeological Method and Theory classes (ANTH 455/555 and 456/556) is the challenge of taking a time-like (materialist) view of the world over a space-like (essentialist) view. Although conceptually we can imagine (or at least pretend to imagine) conceiving of the world where change is continuous and "things" are really events at different temporal scales, our common sense tells us otherwise. Thus, what we see as a "table" is a snapshot (i.e., a particular combination of attributes - or more generally, an event) of a continuously changing set of attributes. Imagine having a time-lapse camera in your head where you can watch an object in "fast forward" mode from the point of its creation to some far distant point. When we do that, the idea of "table" becomes fairly arbitrary because we can view the table from trees to lumber to table to table with marks on it, to wobbly table to refinished table to broken table to firewood to fire (or whatever unique history occurs). The idea of "table" only has meaning as a thing at specific instances in time. This is taking a "time-like" view of the world - and its vital to constructing a scientific understanding of history. Of course, humans are not particularly well-suited to conceiving of the world this way. Our "default" position is that things are real and unchanging. Yeah, we can understand that things change but we are not really good at detecting change. We are good at seeing difference but change is an effect that we largely understand in terms of difference. This has a number of consequences. First, as my colleague Mark Madsen pointed out to me, the cognitive mechanisms that filter information coming from the eyes highlight the irrelevance of the "realism" and "non-realism" debates in the philosophy of science. While there is some "world" out there, we can never escape the set of filters by which we observe it. In addition to all of the culturally mediated (in the Hanson 1958 and Osgood 1951 sense), we can never perceive "reality" since we don't have the mental machinery to do it. The second consequence is that our fallback position (i.e., common sense) is one of essentialism (space like views). If we don't specifically try to build explicit measurement systems (i.e., classification) we will end up making assumptions about thing-ness that embodies space-like perceptions. In other words, we will treat our observations in the here-and-now as if they are the "real" world. The degree to which this aspect of our perception and observation systems is limited by our mental machinery has been demonstrated through a series of experiments by cognitive psychologists. One of the coolest I've seen is a series of videos that demonstrate how poor we are at sensing change. These were done at the Visual Cognition Lab at the University of Illinois.
Friday, March 9, 2007
At the High Table once again
While Mark was down here in Long Beach we decided to finally invest some energy and money into getting the giant pile of RCD videos digitized and on DVDs. This pile of videos, a product of our mid-90s effort to document Dunnell's classes for posterity, represents several of his most influential classes: Archy 497, 498, 479 and 575. Perhaps of interest to only a small few, the videos capture the essence of what was amazing about graduate school in the University of Washington during this time: a somewhat cantankerous man logically constructing archaeology from first principles to sweeping explanations. Although initially baffling, the lessons learned from those classes provided immense metalevel and philosophical understanding of the practice of science and its conceptual framework. Although there are those who really disliked the class, there are probably few that weren't impacted by it in a deep way. Those were amazing classes. Well, to mark the movement of RCD into the digital world, I am happy to announce web versions of Dunnell. I'll try to get more up as they are finished. But for now, treat yourself to the first two lectures of Archy 497: Formal Theory. http://darwin.anth.csulb.edu/rcd/497/Lectures.html
Friday, March 2, 2007
It is really hard to believe the number of these fortified deposits until one goes and actually looks. This one is in Craighead County, Arkansas: 3CG41. I don't think I've seen a deposit as clearly delineated as this one. Absolutely amazing.
Prehistoric Rectangular Fortified Village of the Day...
Ah, Friday. There are few things like the sense of anticipation that arrives on a Friday afternoon. And to celebrate Fridayness, I present another instance of a prehistoric rectangular fortified village deposit. This one is 3MS18 located in Arkansas.
Thursday, March 1, 2007
More fortified prehistoric villages in the LMV
Although we had long worked in this area, it wasn't until we got this photo that it became clear that Caruthersville, MO has a substantial rectangular and fortified deposit. The rectangular feature is clearly visible in this photo.
A couple of minor updates to the seriation macros. These are uploaded at: http://www.csub.edu/~clipo/seriation.html
Subscribe to: Posts (Atom)