This thesis investigates Swedish lexical blends. A lexical blend is defined as the concatenation of two words, where at least one word has been reduced. Lexical blends are approached from two perspectives. First, the thesis investigates lexical blends as they appear in the Swedish language. It is found that there is a significant statistical relationship between the two source words in terms of orthographic, phonemic and syllabic length and frequency in a reference corpus. Furthermore, some uncommon lexical blends created from pronouns and interjections are described. A description of lexical blends through semantic construction and similarity to other word formation processes are also described. Secondly, the thesis develops a model which predicts source words of lexical blends. To predict the source words a logistic regression model is used. The evaluation shows that using a ranking approach, the correct source words are the highest ranking word pair in 32.2% of the cases. In the top 10 ranking word pairs, the correct word pair is found in 60.6% of the cases. The results are lower than in previous studies, but the number of blends used is also smaller. It is shown that lexical blends which overlap are easier to predict than lexical blends which do not overlap. Using feature ablation, it is shown that semantic and frequency related features have the most important for the prediction of source words.

I denna uppsats undersöks franska mots-valises, motsvarande teleskopord på svenska, och deras betydelse. Ett teleskopord består av två eller flera ord som slagits samman. Orden som utgör teleskopordet, les mots sources (källorden) har på något sätt blivit modifierade. Oftast genom att korta av början eller slutet på orden och foga samman dessa. I denna uppsats undersöks hur betydelsen är fördelad mellan de två orden som utgör teleskopordet. Med hjälp av ett hyponymtest utläses vilket eller vilka av källorden som bidrar mest till teleskopordets totala betydelse (ordets ”semantiska huvud”, la tête sémantique). Resultatet visar att det högra ordet ofta fungerar som semantiskt huvud. Eventuella kopplingar mellan källordens form och semantiska bidrag undersöks. Resultatet visar inte på något sådant samband.

<p>Ova disertacija se bavi procesom<br />&nbsp;morfemizacije krnjih osnova koje su nastale<br />od leksičkihslivenica. Rekurzivinim<br />slivanjem s različitim inicijalnim elementima,<br />ove osnove se konsoliduju u svom skraćenom<br />obliku ili oblicima, &scaron;to za rezultat ima<br />separaciju novih tvorbenih jedinica. Najbitniji<br />ciljevi istraživanja su sledeći: da se utvrdi<br />teorijski i terminolo&scaron;ki status ovih elementa i<br />odrede kriterijumi za njihovo članstvo u ovoj<br />kategoriji; da se na osnovu tih kriterijuma<br />napravi popis morfemizovanih elemenata uz<br />razvoj metodologije potrebne za njihovu<br />ekscerpciju; da se sačini spisak leksema u<br />kojima se javljaju; da se ustanove formalne<br />varijacije i njihova relativna učestalost kao i<br />tipovi i oblici inicijalnih elementa sa kojima<br />se kombinuju, &scaron;to bi za rezultat imalo<br />formulisanje prototipskog tvorbenog modela;<br />da istraži značenje ovih elemenata, eventualni<br />razvoj novih semema i smisaonih odnosa; da<br />d&acirc; pregled njihove zastupljenosti u<br />postojećim rečnicima; da predloži model<br />njihove leksikografske obrada i da se na<br />osnovu tog predloga sačini rečnik kako<br />morfemizovanih krnjih osnova, tako i reči u<br />kojima se javljaju.</p> / <p>This dissertation investigates the process of<br />AB morphemization of truncated bases which<br />have originated from lexical blends. Through<br />their recurrent blending with various initial<br />elements, they become consolidated in their<br />truncated form or forms, resulting in the<br />secretion of new word-formational units. The<br />most important goals of the research are the<br />following: to determine the theoretical and<br />terminological status of these elements and<br />establish the criteria for category<br />membership; to compile a list of<br />morphemized elements based on these<br />criteria by developing the methodology for<br />their extraction; to compile a list of lexemes<br />containing these newly morphemized<br />elements; to determine formal variations and<br />their relative frequency as well as the types<br />and forms of initial elements with which they<br />combine, resulting in theformulation of the<br />prototypical word-formation pattern; to<br />investigate the meaning of these elements and<br />the development of new senses and sense<br />relations; to give an overview of their<br />representation in existing dictionaries; to<br />propose a model for their lexicographic<br />treatment and to apply that model in the<br />compilation of a specialized dictionary of<br />morphemized truncated bases and words in<br />which they appear.</p>