Wikipedia:Reference desk/Archives/Language/2014 April 1

= April 1 =

Eastern European naming conventions
Eastern European naming conventions dictate that a female's surname would end in -ska or -ova, and a male's surname would end in -ski or -ov. Now, suppose we have a female, Miss Jakubowska, and she marries an American man named Mr. Smith. By American conventions, she would be known as Mrs. Smith and her children with the man will be known as __________ Smith. Now, if the same person had been male, then he would be known as Mr. Jakubowski, and his children with his American wife, would be... um, this gets a bit tricky. Would the children typically follow Polish naming conventions or American naming conventions? How about a case where a single Polish woman adopts a male child. Would he receive her exact surname (Jakubowska) or the male equivalent (Jakubowski) or his birth name? What if the said single Polish woman is impregnated by a man but does not marry him? Would the child receive the man's surname if the man is known or the woman's surname (masculine version) if the man is unknown? 140.254.227.101 (talk) 17:43, 1 April 2014 (UTC)


 * First of all, your generalization about surname forms does not apply to all cases. It doesn't apply in Eastern European countries with non-Slavic languages, and even in the Slavic countries, not every surname ends in -ski or -ov in the masculine form. Also, the spelling of those forms varies among Slavic countries.  For example, in the Czech Republic, names end in -sky rather than -ski. I can't comment on how Poles handle surnames for single mothers in various scenarios, but I think I can answer the question about the children of Mr. Jakubowski and his American wife.  Assuming that the couple did not adopt a shared surname such as Jakubowski-Smith, then the children would normally take the father's surname.  The form of that surname would probably depend on the country where the couple intended to raise the children.  In the United States, children of Slavic immigrants generally use the masculine form of the father's surname, even if they are girls.  In Poland, I expect that the children's surname would take the appropriate form for their genders.  Marco polo (talk) 19:33, 1 April 2014 (UTC)


 * As for the last question, it's very rare (to my knowledge) for a single woman to give her child the surname of a man she had never married, unless they were in a de facto relationship akin to marriage. Circumstances vary, but generally it would raise too many uncomfortable questions. Or, rather, simple and innocent questions requiring uncomfortable answers.  --   Jack of Oz   [pleasantries]  21:20, 1 April 2014 (UTC)


 * In a country where "gendered" family names are used, the child will always have a properly gendered last name. So a male child of a single mother normally gets her last name, but in masculine form. If you tried to give him a feminine last name intentionally, you probably legally can't, or would find it very hard. Of course, in the US or other places which do not recognize gendered last names, the automatic assumption is that the child just get his mother's name. And you get people with names like Ed Lazowska which seems quite weird to me. --Ornil (talk) 21:57, 1 April 2014 (UTC)


 * It varies in the U.S. some, but quite often the child is given the father's surname even if the mother has no intention of maintaining a long-term relationship with the father. I would say that in the U.S., while far from universal, such a situation is somewhat more common than not.  -- Jayron  32  23:47, 1 April 2014 (UTC)

140.254.227.101 -- I have no idea if it happens in reality, but in Greg Bear's Eon, there's a male U.S. character who has a Slavic feminine last name due to ancestral family history... AnonMoos (talk) 08:55, 2 April 2014 (UTC)


 * In Poland naming conventions are regulated by law. Generally, there are three possible cases, depending on the surname's grammatical characteristics:
 * Name declined like a noun – same form for men and women, e.g.:
 * Jan Nowak, Łukasz Kowal, Jakub Środa (men);
 * Anna Nowak, Olga Kowal, Barbara Środa (women)
 * Name declined like an adjective that ends, in the masculine form, in -ski, cki or -dzki – masculine form for men, feminine form for women, e.g.:
 * Jan Nowacki, Łukasz Kowalski, Jakub Śródzki;
 * Anna Nowacka, Olga Kowalska, Barbara Śródzka.
 * Name declined like an adjective that doesn't end, in the masculine form, in -ski, cki or -dzki – if you're getting married and taking your spouse's surname, you may choose whether you want your new surname's grammatical gender to match your sex or not (the same if you're registering your newborn baby). So if a woman is getting married to a man whose name is Biały, Cichy or Silny, she may choose to be called either Biały, Cichy, Silny or Biała, Cicha, Silna.
 * Of course, these conventions do not apply outside Poland. So if an American woman married a Polish man by the name Kowalski in the U.S., she would likely adopt the name Kowalski. If she then moved to Poland and became a Polish citizen or resident, her surname would not be automatically changed to Kowalska. If she wanted to change her name to a grammatically correct one, she would have to legally request such a change.
 * Sources:
 * — Kpalion(talk) 17:07, 2 April 2014 (UTC)
 * — Kpalion(talk) 17:07, 2 April 2014 (UTC)

Word request
What word would best fit this case? A researcher decides to do a study on whatever, and after collecting data and performing a statistical analysis on the data, he concludes that he's not rejected the null hypothesis and that there is no significant correlation between X and Y at .01 alpha-level of significance. But he further changes the .01 to a .05, giving him more power, and justifies the change. At this point, he declares that he's rejected the null and that there is a significant correlation between X and Y at .05 alpha-level of significance. Obviously, this is academic dishonesty... or perhaps it's pseudoscience? 140.254.226.242 (talk) 20:16, 1 April 2014 (UTC)


 * "If at first you don't succeed, redefine success to include whatever you did." StuRat (talk) 20:34, 1 April 2014 (UTC)
 * On the other hand, "Artificial intelligence is that which hasn't been done." —Tamfang (talk) 05:05, 2 April 2014 (UTC)
 * Tamfang -- "What's easy is hard and what's hard is easy" is a valid maxim in artificial intelligence. Things that are very hard for humans (e.g. long laborious mathematical calculations) are easy for computers, while things that are very easy for humans (vision, motion, language) are extremely difficult to implement in machines.  In the 1960s, before this was fully understood, some people imagined that the development of such things as a computer program which could beat many people playing checkers (draughts) meant that big breakthroughs in achieving artificial intelligence were just around the corner, but that wasn't the case... AnonMoos (talk) 08:39, 2 April 2014 (UTC)


 * I don't know whether there's a word. It reminds me of stories about someone who shot a gun and then drew a target around the bullet hole. &mdash;JerryFriedman  (Talk) 21:56, 1 April 2014 (UTC)
 * On which subject, of course, we have an article. AndrewWTaylor (talk) 18:01, 2 April 2014 (UTC)
 * Pseudoscience can be a lot of things, but generally it doesn't include deliberately falsifying results. So I wouldn't use that word. The best I can find is "scientific misconduct." See the Wikipedia article for Scientific misconduct.--Dreamahighway (talk) 23:41, 1 April 2014 (UTC)


 * It's not dishonest at all. The fact that he was expecting 0.01 doesn't change the fact that the correlation is significant at the 0.05 level, and there's nothing wrong with reporting p=0.03 (for example) if that's actually the case.  Scientists have to deal with unexpected sources of error all the time. It's frequently the case that their results are weaker than what they were hoping for, but that doesn't make the result any less valid.  --Bowlhover (talk) 02:30, 2 April 2014 (UTC)


 * 140.254.226.242 -- Often such "significance" testing isn't what's most practically important anyway, since these numbers have very little to do with effect size. Something can be significant to the .00001 level and still have semi-negligible effect size... AnonMoos (talk) 08:27, 2 April 2014 (UTC)


 * Academic fraud. See Academic dishonesty. Sleigh (talk) 09:39, 2 April 2014 (UTC)


 * It's certainly not fraud or pseudoscience, or even academic dishonesty if the researcher reports the actual data, because then any statistician can make an appropriate inference about statistical significance. A scrupulously honest researcher would report that he was originally intending to test at the 1% level and that the results obtained are not significant at this level, but he could then perfectly justifiably go on explain that the results are significant at the 5% level and that further research might be called for.  Testing of medicinal drugs in the past has often been unreported when the required significance level has not been reached, but regulations in the UK are now requiring all results to be published, even the ones that don't turn out as expected or hoped.    D b f i r s   18:16, 2 April 2014 (UTC)
 * YES to this! About the worst we could call it is "lowering the bar" or some such. As long as the appropriate statistics are reported, there is no misconduct. SemanticMantis (talk) 14:55, 3 April 2014 (UTC)