<section data-background="img/puzzleworld.jpg"> <section style="background-color:rgba(255, 255, 255, 0.5)"> <h1>Linguistics 120: Morphology</h1> <h3>Dr. Will Styler - Spring 2019</h3> </section> --- ### Administrative Notes - My Friday Office Hours are cancelled - But I'll have an extra hour today, after class - I won't be able to respond to emails over the weekend - Amber will give a guest talk Friday - Clickers will be used! - Clicker scores will be finalized early next week along with SONA - If I've asked you to remind me to change anything, now's the time - Graduating seniors, come see me after class! --- ### Grades - There's a good chance that I'll be doing a bit of curving (by HW or overall) - Don't be too disheartened if your current grade is feeling low - Bad grades will still be possible, but I want good students to get good grades - The goal of difficult assignments is to challenge you, not destroy you --- ### Homework 4 is due Friday - **You'll need to work alone on this** --- ### Please fill out your CAPE evaluations - It makes a difference, for me and future students - Be honest, but also be fair --- ### Today's Plan - What more awaits you in Morphology - What we have learned - The secret to great linguistic analysis --- # The Morphology we missed --- ### Quarters are short - There's always a balancing act - "If I focus on this, I'll lose that" - "... but I want to talk about case!" - More time with data is good - ... but takes away time for theory - Morphology is so broad in interactions that you can't cover it all! --- ### There remain many topics for you in morphology! - (We had to save something for grad school, right?) - So, I'll give a sneak preview of some of the things we've missed --- ## Compositionality --- ### Compositionality - "Every complex expression can be exhaustively broken down into minimally meaningful pieces, with no units of form or meaning left unaccounted for” - Are there meanings that are 'unaccounted for'? - Are there words which have more forms than meanings? - Are there words which have more meanings than forms? - Are there words whose meaning cannot be predicted from the forms? --- ### Nahuatl to the rescuetl <img class="r-stretch wide" src="morphology/nahuatlnoncomp.jpg" alt="nahuatlnoncomp.jpg The user wants a detailed description of an image containing text. The image is a list of numbered sentences in two columns. Left column: Sentences in a specific language (likely a Native American language given the orthography like 'kʷahtli', 'tepe:tł', etc., which looks like Nahuatl or a related Uto-Aztecan language, but I should stick to describing what is visible rather than guessing the language unless certain). Right column: English translations in single quotes. Let's break down each line. Line 51: Left: namah mariah ki?ita kʷahtɬi (Note: there are diacritics. The 't' has a macron over it, the 'l' looks like an inverted breve or similar mark, actually looking closer at crop 1 and 2, it's a specific orthography. Let's look really closely at the characters.) - namah mariah ki?ita kʷahtɬi -> Wait, let me re-examine the diacritics. - The 't' in 'kʷahtli' has a macron (line over it). No, wait, looking at crop 1, it's kʷahtɬi. The character after t looks like an l with a breve or similar. Actually, let's look at the full image again. - It says kʷahtɬi. Wait, no. Let's look at line 52: kʷahtemo. Line 53: masatł tʃoloa:. - Okay, I need to be precise with the characters. - Line 51 Left: namah mariah ki?ita kʷahtɬi -> Actually, looking at the very top crop, it looks like kʷahtɬi. Wait, let me look at the full image again. It says kʷahtli with a macron over the t and an inverted breve over the l? No, that's not right. - Let's look at line 53: masatł. That is a 't' with a cedilla or similar below it? No, it looks like a 't' with a horizontal bar below (a tie) or maybe a specific phonetic symbol. Actually, looking closely at crop 2, it looks like t with a breve underneath? Or maybe a specific character from the IPA or a specific orthography. Let's look at line 57: ʃotʃitʃi:wa. - Okay, let's just transcribe exactly what I see as best as possible without over-interpreting obscure linguistics unless it's clear. - Line 51 Left: namah mariah ki?ita kʷahtɬi -> Wait, looking at the full image, the last word is kʷahtli with a macron over the t and an inverted breve over the l? No, let's look at line 52. kʷahtemo. - Let's try to read the diacritics carefully. - Line 51: namah mariah ki?ita kʷahtɬi -> The last character looks like an 'l' with a breve over it? Or maybe a specific symbol. Actually, looking at line 53 masatł, that is a 't' with a cedilla? No, it's a 't' with a horizontal bar below. Wait, let me look at the high-res crop if possible. I don't have high res. - Let's look at the provided image again. - Line 51: namah mariah ki?ita kʷahtɬi. The last letter looks like an 'l' with a breve over it (like a hook). Or maybe it's just a specific font for 'li'. Wait, looking at line 53 tʃoloa:, there is a colon. - Let's look at the diacritics again. - Line 51: kʷahtɬi. The last character looks like an 'l' with a breve over it? No, looking at line 53 masatł, that is a 't' with a cedilla? No, it's a 't' with a horizontal bar below. Wait, let me look at the very first crop. - Ah, I see kʷahtɬi. The last character is an 'l' with a breve over it? No, that's not standard. Let's look at line 53 again: masatł. That looks like a 't' with a cedilla (comma below). - Actually, let's look at the text kʷahtli in line 51. The 't' has a macron over it. The 'l' has an inverted breve? No, that's not right either. - Let's look at line 53: masatł. That is a 't' with a cedilla below it. And tʃoloa:. The 'o' has a macron over it. - Okay, let's just describe the visual appearance of the text. - Line 51 Left: namah mariah ki?ita kʷahtɬi -> Wait, looking at crop 1, the last word is kʷahtli. The 't' has a macron over it. The 'l' has an inverted breve (like a hook) over it? No, that's weird. - Let's look at line 52: pedroh kʷahtemo tepe:tł. The last word ends in tł. That is a 't' with a cedilla below it. - Line 53: ne masatł tʃoloa:. masatł has a 't' with a cedilla. tʃoloa: has a colon at the end and a macron over the first 'o'. Wait, looking closely at crop 2, it's tʃoloa:. The 'o' has a macron. - Line 54: nokoneh masatʃoloa:. Wait, is that a 't' or 'ch'? It looks like masatʃoloa:. The 't' has a cedilla? No, it looks like masatʃoloa:. Wait, let's look at the full image. - Line 54 Left: nokoneh masatʃoloa:. The word is masatʃoloa:. There is a macron over the 'o'. - Line 55: notatah kioni a:tł imilah. a:tł has a colon and a 't' with a cedilla? Or maybe just a specific symbol. It looks like a:tł. - Line 56: a:onijaja. - Line 57: ʃotʃitʃi:wa. Okay, let's try to be as accurate as possible with the transcription. The text appears to be from a linguistics textbook or language learning material, likely Nahuatl given words like tepetl (mountain), eagle, deer. Actually, looking at the orthography: - kʷahtli = eagle (Nahuatl) - tepe:tł = mountain (Nahuatl) - masatł = deer? No, mazatl is deer. So masatł might be a dialect or specific orthography for deer. Or maybe masa + tł. - tʃoloa: = flee/run away (Nahuatl choloa). So this is Nahuatl. I can mention that context if relevant, but the prompt asks to describe the image. Structure of the description: 1. Overall layout: Two columns, numbered list from 51 to 57. 2. Left column content: The foreign language text with specific diacritics. 3. Right column content: English translations in quotes. 4. Specific details on each line. Let's refine the transcription of the difficult characters. - Line 51: namah mariah ki?ita kʷahtɬi. Wait, looking at crop 1, the last letter is an 'l' with a breve over it? No, that looks like a specific symbol for a lateral fricative or something. Actually, in Nahuatl orthography, tl is often written as such. But here it looks like kʷahtli. The 't' has a macron. The 'l' has an inverted breve? No, let's look at line 53 masatł. That is definitely a 't' with a cedilla (comma below). - Let's look at line 51 again. kʷahtli. The 't' has a macron. The 'l'... actually, looking really closely at the full image, it looks like kʷahtɬi where the last character is an 'l' with a breve over it? No, that's not standard. - Let's look at line 52: pedroh kʷahtemo tepe:tł. The last word ends in tł. That is a 't' with a cedilla below. - Line 53: ne masatł tʃoloa:. masatł has a 't' with a cedilla. tʃoloa: has a colon at the end and a macron over the 'o'. Wait, looking at crop 2, it's tʃoloa:. The 'o' has a macron. - Line 54: nokoneh masatʃoloa:. Wait, is that a 't' or part of the digraph? It looks like masatʃoloa:. There is a macron over the 'o'. - Line 55: notatah kioni a:tł imilah. a:tł has a colon and a 't' with a cedilla? Or maybe just a specific symbol. It looks like a:tł. - Line 56: a:onijaja. - Line 57: ʃotʃitʃi:wa. Let's look at the diacritics again very carefully. - Line 51: kʷahtɬi -> The last character looks like an 'l' with a breve over it? No, looking at crop 1, it looks like kʷahtli. The 't' has a macron. The 'l' has... actually, it looks like a specific symbol. Let's assume it's kʷahtli with diacritics. - Actually, let's look at the provided text in the prompt's image again. - Line 51: namah mariah ki?ita kʷahtɬi. The last character is an 'l' with a breve over it? No, that looks like a specific symbol for a lateral fricative /ɬ/. Yes! In Nahuatl orthography (specifically the one used by linguists), tl is often written as such. But here it looks like kʷahtli. Wait, looking at crop 1, the last letter This description was generated automatically. Please feel free to ask questions if you have further questions about the nature of the image or its meaning within the presentation."> --- ### It sure seems like there's non-compositional meaning! - 'Drink-water' becomes 'drown' - ... but only when incorporated! - 'Flower-make' becomes 'arrange flowers' - ... but only when incorporated! - 'deer-flee' becomes 'skittishly and quickly run away' --- ### "... but maybe those are just morphemes with different meanings!" - Well, OK, but how about... --- ## Phonaesthemes --- <section data-background-color=#708cef></section> ### "Would you like to buy a new Glimp? It..." a) Sharpens knives. b) Lights up your house. c) Tastes good. --- <section data-background-color=#708cef></section> ### "Damnit, I left my meb in the car!" a) I won't be able to sew as well. b) I'll be unable to fill my bird feeders c) What if there's a fire?! --- <section data-background-color=#708cef></section> ### "Have you seen my snofter?..." a) My nose is sore. b) I dropped a glass. c) It's cold in here. --- <section data-background-color=#708cef></section> ### "I got diabetes. It must be from drinking all that..." a) Blurge cola b) Mabler cola c) Wika cola --- <section data-background-color=#708cef></section> ### "You Mom is plafty." a) "Them's fighting words." b) "Maybe?" c) "Thanks!" --- <section data-background-color=#708cef></section> ### "Your Mom is slafty." a) "Them's fighting words." b) "Maybe?" c) "Thanks!" --- ## Phonaesthemes "Frequently recurring sound-meaning pairings that are not clearly contrastive morphemes" - First defined by Ben Bergen ([Bergen 2004](http://muse.jhu.edu/login?auth=0&type=summary&url=/journals/language/v080/80.2bergen.pdf)), who is at UCSD! --- ## Ben Bergen <p>(Audiovisual content has been removed for compliance with recent federal accessibility guidelines. <a href="https://accessibility.ucsd.edu/policies-standards/ucsd-accessibility-guidelines.html">Please see this site for details.</a>)</p> --- ### Phonaesthemes in English - gl- ‘light, vision’ *glimmer, glisten, glitter, gleam, glow, glint*, etc. - sn- ‘nose, mouth’ *snore, snack, snout, snarl, snort, sniff, sneeze*, etc. - sl - 'pejorative', *slack, slouch, sludge, slime, slash, sloppy, slug, sluggard, slattern, slut, slang, sly, slither, slow, sloth*, etc. - b_l - 'expand abnormally', *blimp, bulge, bulk, balloon, blip* --- ### These are clearly not morphemes - In fact, they're part of other morphemes (e.g. "snack") - ... but they sure carry meaning! - These *emerged* as statistical accidents in the lexicon - "Lots of GL words mean light, so GL must mean light!" - They do end up seeming 'real' to speakers! - Neat! --- - Speaking of things that aren't morphemes... --- ## Lessthemes --- ### Nah, that's not really a thing. - ...but for half a second, you all were reanalyzing 'morpheme' pretty hard - <img class="r-stretch wide" src="humorimg/evillaugh.gif" alt="evillaugh.gif This image shows a close-up shot of a man who appears to be the actor Vincent D'Onofrio portraying the character Wilson Fisk, also known as Kingpin, from the television series Daredevil. The man is bald with fair skin that has a shiny or sweaty appearance, particularly on his forehead. He has a large, prominent nose and thick, dark eyebrows that are furrowed in an expression of anger or intensity. His eyes are narrowed and he is looking directly forward. His mouth is slightly open, revealing his teeth, as if he is shouting or speaking aggressively. He is wearing a light grey or beige jacket with a high collar that stands up around his neck. The fabric appears smooth. In the foreground on the left side of the image, there is a hand reaching toward the camera. This hand is out of focus and blurry due to being close to the lens, suggesting he is lunging forward or attacking. The background is dark and indistinct, appearing to be a shadowed wall or room, which keeps the focus entirely on the man's face and upper body. The lighting is dramatic, casting shadows that emphasize his facial features. This description was generated automatically. Please feel free to ask questions if you have further questions about the nature of the image or its meaning within the presentation."> --- Instead, let's check out... --- ## Multi-Word Expressions --- ### We talked about compounds - "Bleu Cheese Salad Dressing" - "Laptop" - "Deer-Flee" - "Inflectional Morphology" --- ### Compounds are continuous chunks! - "I'm going to Bleu Cheese the heck out of this Salad Dressing" is not a compound use - Although "blue cheese" and "salad dressing" may still be --- ### There are other 'words' which are discontinuous - "I threw my lunch up on the roof" - "I threw up my lunch on the roof" - "We're going to cut Mary out of the agreement" - "We're going to cut out Mary from the agreement" --- ### These verb-particle constructions feel like words in many ways - "Hit up" != "Hit" - "Throw out/up/in/back" - "Cut up/across/down/out/in/with" - They're separable, though - ... and comprised of multiple parts - These are super difficult for computers to handle --- ### Speaking of which... --- ## Computational Morphology --- ### Morphology complicates natural language processing - We need some way to teach Siri that 'walks' and 'walk' and 'walking' and 'walker' and 'walked' are the same - "I saw Cardi B's dog walk to the store" and "I saw Cardi B dog walk Tomi Lahren" - "Siri, remind me to hit Jamie up" - "OK, I'll remind you to hit Jamie." - "User has entered a word "swash". What's the plural of it?" --- ### Quantitative Research into Morphology is crucial - How can we solve morphological problems by language modeling - How can we use large amounts of data predict to morphological behavior? - What can huge corpora teach us about usage? - How does machine learning help us understand human learning? --- ### The interface between morphology, learning, and computer work will be Friday - Amber's got a great talk planned - Be there, or be super lame - Speaking of interfaces... --- ## Interfaces --- ### Morphology connects to everything! - Phonetics can be affected by lexical considerations - Morphophonology is 100% a thing - Morphology and Syntax interact at many levels - Semantics is the meaning half of form and meaning - Pragmatics is crucial for context and contrast --- ### There's much more waiting for you in Morphology! - That's why it's called 'More-phology!' - <img class="r-stretch wide" src="humorimg/badpundog.jpg" alt="badpundog.jpg The image shows a close-up of a dog that appears to be a Siberian Husky. The dog has striking blue eyes and a coat with a mix of white and gray fur, which is characteristic of the breed. Its mouth is open wide in what looks like a playful or excited expression, showing its teeth. The dog's ears are perked up, indicating alertness. The dog seems to be interacting with a stuffed toy that resembles another animal, possibly a bear or a similar creature. The toy has a grayish color and appears soft and plush. The background is indoors, with neutral-colored walls and what looks like the edge of a piece of furniture partially visible on the right side of the image. There are no numbers, text, or diagrams present in this image that would require description. This description was generated automatically from image files by a local LLM, and thus, may not be fully accurate. Please feel free to ask questions if you have further questions about the nature of the image or its meaning within the presentation."> --- # Looking back at this quarter --- ### What did you learn about... - Morphological Analysis? - Morphological patterns? - What 'word' means? - How words might be created? - About the lexicon? --- ### You've learned a lot! - You can now solve morphological problems with elegance, generality, realism, and a minimum of complexity - You know how morphology works, around the world! - You know some ways of dealing with weirdly shaped 'words' and affixes! - You've got a good sense of ways too think about word creation and productivity! - You know the big questions about the lexicon, and a few different answers! --- ### We've also had some big discussions about the nature of language - Should we describe language through parts and rules, which *generate* the proper forms? - Should we describe language by looking at the *usage*, and the make generalizations as needed? - Should we take the middle path, and say that grammar may emerge, but the patterns are real, too? - Which of these is cognitively real? - ... does it even matter? --- ### ... and you learned some valuable life lessons - A good analysis isn't good if you can't understand it - Don't assume that anything must be (or mustn't be) marked - Don't assume that 'salt' isn't secretly a verb. - Don't assume that a new language will work like you're used to - Don't open a beer with a chainsaw, just because you can. - **The past tense of 'yeet' is 'yote', and the participle is 'yaughten'** --- ### I know what you're thinking... - "Wow, all that amazing knowledge, in just 10 weeks?" - **"What more could we possibly ask for?"** --- # The secret to great linguistic analysis --- ### You've come a long way in the program - LIGN 101, 110, 111 - Many of you have taken syntax and semantics - Some of you are about to graduate! --- ### In the future, you'll all need to do linguistic analysis - Perhaps as part of your teaching or research, if you decide to go the academic route - Maybe in industry or defense, if that's your path - Maybe for fun, on the weekends - ... and during the week --- ## You'll want to do *great* linguistic analysis - So, I'm going to finally tell you the secret --- ### Most linguistics students don't learn this until graduate school - Many not until a post-doc - Some, not until they're full professors - Some poor souls never figure it out! - **... but you'll know it today!** --- ### We'll end the class like we've lived the class, staring at data! - *To discover the secret to Great Linguistic Analysis, you'll need to figure out...* --- ### What is the Russian marker for 1SG.PRES for these verbs? <style type="text/css"> .tg {border-collapse:collapse;border-spacing:0;} .tg td{font-family:Arial, sans-serif;padding:10px 5px;border-style:solid;border-width:1px;overflow:hidden;word-break:normal;border-color:black;} .tg th{font-family:Arial, sans-serif;font-weight:normal;padding:10px 5px;border-style:solid;border-width:1px;overflow:hidden;word-break:normal;border-color:black;} .tg .tg-0pky{border-color:inherit;text-align:left;vertical-align:top} .tg .tg-0pk{font-weight:bold;border-color:inherit;text-align:left;vertical-align:top} </style> <table class="tg"> <tr> <th class="tg-0pk">Russian</th> <th class="tg-0pk">English</th> <th class="tg-0pk">Russian</th> <th class="tg-0pk">English</th> <th class="tg-0pk">Russian</th> <th class="tg-0pk">English</th> </tr> </b> <tr> <th class="tg-0pky">djelat</th> <th class="tg-0pky">'to do'</th> <th class="tg-0pky">pit</th> <th class="tg-0pky">'to drink'</th> <th class="tg-0pky">staret</th> <th class="tg-0pky">'to grow old'</th> </tr> <tr> <th class="tg-0pky">djelaju</th> <th class="tg-0pky">'I do'</th> <td class="tg-0pky">pju</td> <td class="tg-0pky">'I drink'</td> <th class="tg-0pky">stareju</th> <th class="tg-0pky">'I grow old'</th> </tr> <tr> <td class="tg-0pky">djelajet</td> <td class="tg-0pky">'(s)he does'</td> <td class="tg-0pky">pjet</td> <td class="tg-0pky">'(s)he drinks'</td> <td class="tg-0pky">starejet<br></td> <td class="tg-0pky">'(s)he grows old'</td> </tr> <!-- <tr> <td class="tg-0pky">tʃitaju</td> <td class="tg-0pky">'I read'</td> <td class="tg-0pky">risuju</td> <td class="tg-0pky">'I draw'</td> </tr> <tr> <td class="tg-0pky">tʃitajet</td> <td class="tg-0pky">'(s)he reads'</td> <td class="tg-0pky">risujet</td> <td class="tg-0pky">'(s)he draws'</td> </tr>--> </table> --- ## That's the secret, right there - That morpheme - As you go forward in your linguistic career, remember that... --- # The Secret to Great Linguistic Analysis was /ju/ all along! --- ### For Next Time - Amber will give an awesome talk about her work on French - You'll turn in HW4 Friday - **Graduating Seniors, come see me after class!** --- <huge>Thank you!</huge>