1. Technological discussions in iron and steel, 1871-1885 Carol Siri Johnson, New Jersey Institute of Technology Peter B. Meyer Research Economist, U.S. Bureau of Labor Statistics * SHOT conference, Las Vegas, Oct 13, 2006 * Views expressed here do not reflect official policies or measurements of the U.S. Bureau of Labor Statistics.

2. Mid-19 th Century Ironworks Lukens Steel, circa 1895, just before being torn down Courtesy Hagley Museum and Library

3. Late 19 th Century Ironworks Birmingham, Alabama. Courtesy IA, Robert Gordon, the Smithsonian

4. 20 th Century Ironworks Bethlehem Steel, 1931 Courtesy Hagley Museum and Library

5. Transactions of the American Institute of Mining Engineers (TAIME)

6. The American Institute of Mining Engineers and the Engineering and Mining Journal

7. Knowledge Transfer, the Engineering and Mining Journal and TAIME

8. Online Archive of PDFs at http://techterms.net/ironwork/TAIME/ Vol. I1871-1873Vol. VIII1879-1880 Vol. II1873-1874Vol. IX1880-1881 Vol. III1874-1875Vol. X1881-1882 Vol. IV1875-1876Vol. XI1882-1883 Vol. V1876-1877Vol. XII1883-1884 Vol. VI1877-1878Vol. XIII1884-1885 Vol. VII1878-1879Vol. XIV1885-1886

9. Data Exploration: Paper Length and Contents

10. Data Exploration: Author Data

11. Data Exploration: Author Biographies

12. Kuhn’s hypothesis An established scientific paradigm has a specialized vocabulary Covering its esoteric theory and subject precisely (p. 206) Definitions are established and standard As a paradigm is developing, communication involves translation (p. 203) And exploring alternative definitions (p. 200) (e.g.: “what I mean by steel is... “; also we see glossaries) “The price is often sentences of great length and complexity.” (p. 203) Mass production steel was a developing technological paradigm in 1871-1885 Bessemer steel Open-hearth steel Big Steel – centralized, high volume, capital-intensive production We look at how lengths of articles change in this developing literature. Possibly, articles on a topic get shorter with time.

13. Expressions of literal uncertainty Compute the fraction of words in each article contained in these iron-related phrases: “hot blast”, “Bessemer”, “puddling”, “open hearth”, “Siemens”, “Martin”, “spiegel” This frequency is a proxy for the iron-relevance of the article. Analgously, compute the frequency the string "uncert", representing relevance of literal uncertainty. These counts across the 712 articles have a positive but tiny correlation:.0071.  So articles with iron-related terms are slightly more likely to use words with "uncert" with them than other TAIME articles are. Iron-related terms were declining in frequency in TAIME although the iron sector was expanding as a fraction of the economy at large. Perhaps this was because of the rise of corporate research and development; maybe iron papers were less likely to be published in open publications like TAIME.

14. Word Counts – Original List ? ambiguity answer (s, ed, ing) argument (s) ask (s, ed, ing) belief (s) certain (ty) chance confusion criticism definite difficult dispute (s, un) doubt (s, ful, fullness, less, lessly) error (s) evidence experiment (s, al, ing) if inquire (d, ing) inquiries inquiry investigate (s, ed, ing) know (un, n. ledge) opinion (s) perhaps probability problem (s) query question (s, ed, ing, able) resolve (s, d, ing) right risk (s, ed) test true trust (dis, mis, s) truth (s) (un) certain (ty) valid (ity) whether wrong

15. Word Counts – Final List

16. Expressions of Uncertainty

17. Scientific Expressions

18. Expressions of Uncertainty – Steel Articles

19. Scientific Expressions – Steel Articles

20. Layers of productive processes, advancing Iron and steel, 1871-1884 Blast furnaces making pig iron Bessemer and open hearth steel production Iron and steel plants Railroads (transportation) Business process of railroad companies (cost accounting, personnel departments, time setting, timekeeping) Earlier, more basic, "upstream" levels Materials science and solid state physics Chip design and electrical engineering Semiconductor memory and microprocessor chips Microcomputers Applications software (word processors, spreadsheets, databases, chip design software) Net software and business process (e-commerce, auctions, search engines) Production of information technology goods, in recent decades Later, “downstream" levels In both cases there were feedback processes by which downstream advances affected earlier stages of production