tag:blogger.com,1999:blog-8386055846297828307.post2611575171251349217..comments2024-03-28T07:34:49.133+01:00Comments on The Genealogical World of Phylogenetic Networks: Typology of semantic change (Open problems in computational diversity linguistics 8)Unknownnoreply@blogger.comBlogger2125tag:blogger.com,1999:blog-8386055846297828307.post-47756979686864534122019-10-31T18:43:28.274+01:002019-10-31T18:43:28.274+01:00Yes, it's again the system, I completely agree...Yes, it's again the system, I completely agree. We won't be able to avoid it, but we can, hopefully, isolate it in such a way that we could say: if the system is not active, a change from a -> b is more likely than a -> c. But you are completely right. It doesn't stop to be complicated...Johann-Mattis Listhttps://www.blogger.com/profile/12185355133399020406noreply@blogger.comtag:blogger.com,1999:blog-8386055846297828307.post-9049261451858779132019-10-16T20:27:37.451+02:002019-10-16T20:27:37.451+02:00This may be getting ahead of things, but I would s...This may be getting ahead of things, but I would suggest that a fourth problem is posed by metricization. Once we have data about what happened where, how do we actually compare particular semantic changes with one another?<br /><br />The simplest option would be a vaguely Swadeshian one-dimensional measure: "change X happens over Y years with a Z% probability". This is likely too simple though. Semantic change can be probably "conditioned" by various factors, and quite likely we would end up seeing areal or genetic patterns in how often a given change occurs where.<br /><br />Two proposed factors that seem rather likely to me are "polysemy pressure" and "synonymy pressure" (discussed by George Starostin e.g. in his 2010 article on defining "preliminary lexicostatistics"). But these are factors that cannot be themselves measured in isolation! They require examining the full lexicon of the language. Essentially this would then mean that every semantic change happens within a particular unique semantic ecosystem, and is therefore a unique event, not exactly compareable with any other semantic change (even if its input and output are the same). So if we want to know e.g. if a change from 'warm' to 'hot' is more probable than the opposite? We could end up only being able to say that the former is known to happen under conditions A B C and D, while the latter is known to happen under conditions E and F, and unable to offer any typological argument on what would be likely to happen under conditions G or H.J Pystynenhttp://protouralic.wordpress.com/noreply@blogger.com