Lexical Bundles in Academic Writing: Advanced Analysis

[00:00:03][Music] Cortez, extending the analysis of lexical bundles in academic writing.

[00:00:23][Music] Viviana Cortez received her PhD in applied linguistics from Northern Arizona University in 2002 and she worked for the program of applied linguistics and teaching English as a second language at our AWA state university for six years.

[00:00:56]Iowa, sorry, uh for six years before coming to Georgia State in the fall of 2008.

[00:01:04]At Iowa State, she taught descriptive grammar, discourse analysis, English for specific purposes and advanced academic writing for international graduate students. Before coming to the United States to pursue her graduate studies, she worked as an English as a foreign language teacher in Buenosiris, Argentina, her hometown, in different elementary and high schools, as well as in a teaching training program. Her research interests include the analysis of recurrent word combinations such as lexical bundles in different academic registers and different types of corpus-based grammatical studies as well as the use of corpora in teach in the teaching of academic writing. She has presented her research at numerous local conferences and in conferences and symposium in Spain, Mexico, Argentina and other countries. her her research articles can be found in journals such as English for specific purpose, applied linguistics, the journal of English for academic purposes, linguistics and education and corporate and in several edited books. So, it's it's our honor, it's my pleasure to have Dr. Viviana Cortez is here with us in a partnership with the regional English language office RLO and the American Embassy.

[00:02:38]U welcome uh Professor Vivana Cortis.

[00:02:43]>> Okay, here I am. Um I thank you so much for the invitation um to give my talk um virtually to this audience. Um if we can show the uh presentation PowerPoint if we can share that I would like to start here it's coming here it is okay so for today's presentation um I want to talk about um a special type of formulaic language that I have been investigating for the last 25 years. Uh in fact I was um a doctoral student when I discovered these expressions called lexical bundles and um I have always been very interested in academic writing as a genre in any language. My first language Spanish and my second language English particularly.

[00:03:49]Erh so I started looking at formulaic language many many years ago and I want to share some um some things that I've been doing um in these past 25 years. First of all someone's going to take a picture of the screen. So um um if you have a camera and you are on the screen uh you may be find yourself in some English language program site.

[00:04:16]This is what I uh prepare for today.

[00:04:18]This is the outline. I want to talk about what was the impetus for this study that I am going to present. I'm going to describe my observation which are lexical bundles. Um they are considered building blocks and I want to explain why. Um I want to speak a little bit about the identification of lexical bundles particularly in different sections of the research article. I want to talk about corpus and uh corpus analysis and methodologies that we use to identify and analyze bundles. Then I want to talk a little bit about some studies that I conducted about 10 years ago in which I tried to investigate the different lexical bundles that I used h in moves and steps of research article sections.

[00:05:15]Um I want to go on extending the analysis of lexical bundles particularly the functions of the bundles um to talk a little bit about um how bundles can help convey communicative purposes which is so important when you are writing academically and I'm going to give you some ideas for pedagogical applications.

[00:05:43]S so um when I started looking at lexical bundles a long long time ago okay 25 years ago lexical bundles were so innovative um that we didn't even have commercially available computer programs to identify bundles um formulate language and collocations because after all bundles extended collocations.

[00:06:15]They have been investigated for many many years for decades, but they were usually investigated from a perceptual point of view because there was no way to identify them empirically.

[00:06:31]Before computers and before com corpora, um we had intuitions about which expressions repeated frequently uh which uh words collocated frequently with other words.

[00:06:47]But when computers started to be used for linguistic analysis, we had more resources to um identify collocations uh words that prefer to be with some words instead of others. So when I was a PhD student, I had to create my own computer program that I call the LBP, the lexical bundles program. And I had to learn how to program and make my own program.

[00:07:16]Nowadays um there are computer programs that are used uh that you can get online that you there's software that is freeware that you can get um to um analyze um identify and analyze lexical bundles. But in those days everything was new and everything we did when we identify bundles was to classify them structurally and then classify them functionally. And for many years um ever since uh Biver and his collaborators coined the term lexical bundles to refer to these expressions. Um you can find studies in the top journals in applied linguistics that did exactly that. They identified lexical bundles. They analyze and classify them structurally and then they used some already designed taxonomy to classify them functionally. There are dozens of um erh articles that were published every year from the 2000 till 2020 that do exactly that.

[00:08:34]Now when I started my first studies were exactly like that. I was collecting a corpus, running my program, getting the bundles, analyzing them structurally, then functionally and that was it. Then I realized that the functions of the bundles are not as simple as I thought at the very beginning. that these bundles are really building blocks that help the writer or the speaker. In my case, the writer because I um I my area of specialization is um academic writing. So the these bundles help the writer um with their communicative purposes, their communicative intentions, how to convey a certain communicative uh purpose.

[00:09:33]So um I realized that the more specific the type of texts you are analyzing, the more specific is the function of the bundle. And that was something that I um realized was happening because I had a very large corpus of research articles and then I decided to separate that corpus into sections. So I had a subcorpus of introductions, a sub subcorpus of um methods, a subcorpus of results, a subcorpus of discussions. And when I identify bundles in those subsections, the functions were much more specific.

[00:10:23]I have a lot of lists of bundles, many many because um having a program that identify the bundles that is very reliable and um having discovered these corpusdriven approach to identify bundles um was a very powerful tool. I could think of researchers um in the 40s or 50s like FTH for example who is like the father of collocations that not even in his wildest dreams he could have imagined that someone could have a program that identifies empirically which are the most frequent collocations in a in a corpus in in a group of texts.

[00:11:10]So to me because I am a a teacher a writing teacher I taught academic writing to international graduate students for many years in the United States.

[00:11:22]For me it's always very important to think of the pedagogical application of my research what we call in the states the scholarship of research and teaching. Okay. So um you um you think of doing research that has a very important pedagogical impact.

[00:11:43]So when I do studies of lexical bundles I always think of how this is going to help my students write better or be more successful with their writing. So I always think of ways in which I can teach these lexical bundles.

[00:12:03]So for those of you who have never heard or seen lexical bundles, let me give you some definitions and some examples. So lexical bundles are by definition sequences of three or more words that commonly go together in natural discourse. Okay. They had identified by a specially designed computer program.

[00:12:27]This program is a program that looks at the first word of a text in a corpus.

[00:12:34]You know, a corpus is a collection of texts. So the the program looks at the first word of the first text and moving one word at a time identifies an expression. For example, I want to identify four word bundles. So my program will look at word number one, word number two, word number three and word number four. We consider that a for a database.

[00:13:09]Then it will move one word to the right and would look at word number two, word number three, word number four, word number five. will get those four words.

[00:13:21]That is going to be a foreground and it's going to put it in a database.

[00:13:25]Every time the program does that, it's going to look at the database to see if that those four words are already in the database. If they are already in the database, the program is going to flag that and it's going to add one more count. The program moves like that until it finds a punctuation mark. When the punctuation mark is found, the program stops and goes again. It starts counting one, two, three, four after the punctuation mark. Um, lexical bundles don't have punctuation marks because they never cross clauses or sentences.

[00:14:08]um when it gets to the end of a file um of a text, the program goes to the next text. And so it does the same procedure counting one, two, three, four. But this time when it looks at the database, it doesn't only look to see if it was there before. It only looks to see if it was in a different file. Because the important thing with bundles is the frequency and the range. So a word combination must recur very frequently to be considered a lexical bundle. So for example, for expressions that have four words that that that is four word lexical bundles, they have to repeat 20 times in a million words and at least in five texts and sometimes in more texts depending on how many texts your corpus has. This is called range and it's a measure that ensures that you are not looking at an expression that is being used by one user or a very few users of the language.

[00:15:27]In applied linguistics, lexical bundles are considered building blocks that are used by speakers and writers in a register. They are considered building blocks because they are these linking expressions that can help connect um expressions that are used forformational purposes. And in case you have never seen a lexical bundle, some examples of lexical bundles are expressions like as a result of, on the other hand, in the context of, as well as there. to mention just a few because there are many many these are four word lexical bundles.

[00:16:09]Now one thing that you need to consider is that lexical bundles are not complete units. Okay, most lexical bundles are made up of fragments of expressions.

[00:16:25]um they are classified usually by the correlates that they um connect to. For example, as a result of would be a phrasal bundle because it starts with a prepositional phrase. It starts with a preposition as. And so you say that this is a prepositional bundle, but it is made up of a prepositional phrase that is complete as a result, but it has the fragment of another prepositional phrase embedded. That would be the preposition that starts with of. So for example, in the context of you have another one there that is a prepositional phrase that has the fragment of another prepositional phrase that it starts with of but it doesn't finish.

[00:17:17]Now many years ago because of this um conversation that I had with my major professor Doug Bber that we realized that lexical bundles have a more specific function in more specific discourse. I decided to create a corpus of research article introductions. I had a corpus of research articles that was very very large. Okay, probably four or five million words of research articles in many disciplines. So I wanted to conduct a study just to see if this idea that Bible and I had had that the more specific the corpus, the more specific the function could be empirically proven. So I got 1,200 research article introductions because you know introductions are short. So in order to get to a million words, which is more or less the threshold that you need to identify bundles, I needed a lot of introductions.

[00:18:23]So I um collected these, 1200 introductions from 13 different disciplines. My purpose was not to um conduct any interdisciplinary study but I thought that to have a corpus more or less representative of different disciplines I was going to collect the corpus from uh introduction from different disciplines. So there you have the list of disciplines that um the introductions that I work with came from. Grononomy, applied linguistics, animal science, biology, business, chemistry, civil and materials engineering, communication studies, computer science, economics, physics and astronomy, um statistics and urban planning. So a little bit of everything.

[00:19:16]I still had my lexical bundles program that I had created when I was a student.

[00:19:21]But when I created the study for my dissertation, I was only looking at four word bundles. So my program only looked for four word bundles. However, for this study, I wanted to see if I could identify longer bundles as long as I could get one. And so I twitched the program a little bit so that I could create um I could identify bundles that were four, five, six, seven, eight as long as possible. The longest I found was nine words.

[00:20:03]And then following the tradition that I had been following since I was a student, I identified the bundles. I classified them structurally and then I classified them functionally. I used um the taxonomy that I had created together with my teacher that piper and another colleague of ours Susan Cornra. We have done that taxonomy for an article that we published in apply linguistics in 2004. And it's a very basic taxonomy, but it's always good to use that taxonomy to start. Let's see if I have some water here.

[00:20:45]I had some water, but now we don't have any water. Okay, I apologize for this cold that I have.

[00:20:54]It's not letting me speak very clearly.

[00:20:57]Um now I realized that lexical pandles go hand inhand with many moves and steps as those identified by John Swales in his studies on rhetorical moves in research article introductions.

[00:21:18]So in an article that was published in uh 2013, I identified those bundles. I studied those bundles that I had identified with my newly refurbished program and I tried to see if I could connect the bundle and the move. And I realized that there are many bundles of four, five, six, 7, 8, 9, 10, nine words up to nine words that connect to a move and sometimes to a move and a step inside the move directly.

[00:22:04]And I call those trigger bundles because those bundles trigger the move. So I identified 135 bundles and 35 out of those 135 were trigger bundles and they were connected to one move. For example, the bundle the purpose of this study is to was used exclusively in move three step one announcing present present research purposively. So here you have an example from my corpus from business a business research article that starts with that bundle. Move three starts with the purpose of this study is to develop and test a model that integrates various socioeconomic factors including culture, public bureaucracy and economic conditions. Another bundle a better understanding of was connected to move to establishing a niche presenting positive justification. So here you have an example from biology that says a better understanding of organ formation would thus profit from a tempor temporally and especially restricted target to avoid blurring of differences in RNA expression levels. So by introducing that bundle the author was establishing the niche by presenting a positive justification of a procedure used in biology.

[00:23:48]Now for example here you have move one establishing a territory. This is from Ses 1981 2004 because he created the first test scheme in 1981 but he re um refined it in his work on research genres in 2004 and he has this move one establishing a territory that has three steps. Step one, claiming relevance of the field. Step two, making topic generalizations. And step three, reviewing items of previous research.

[00:24:29]And these bundles that you can see on the second column were found to be directly connected to those steps. So when you claim relevance authors where you think one of the most important one of the most play an important role in one of the most for some superlative the importance of the in the case of making topic generalizations they were using it is well known that there has been a it has been shown um things have been shown too and then reviewing items of previously literature. It has been suggested that in a number of studies have shown that it was found that.

[00:25:19]However, I as I just said only 35 bundles were trigger bundles. The rest of the bundles were non-trigger, meaning they could be connected to a particular move and step.

[00:25:35]They were more of a wild card. They could work in more than one move or in more than one step because they were not used to trigger the move. So trigger bundles are easily identified in the move. However, not trigger or at the very beginning I used to call them comment because it was as if you were commenting making a commentary on whatever was triggering the move. So these non-trigger bundles were related to particular moves but it was more difficult to know when to use them. For example, look at this. This is an example from business. And as you can see the first bundle little is known about is showing the gap. Okay. The problem is little is known about the behavior of acrual in the absence of bias but in the absence of of it's also a bundle but it's used across more than one move and it's not directly triggering the move. So the idea is when you teach lexical bundles, if you are teaching a move and a step scheme to your students, you can it's very easy to teach the bundles that trigger the move.

[00:27:02]But what do you do with the bundles that don't trigger the moves because those ones are more difficult to identify and the function is more difficult to define.

[00:27:14]So the idea is you have to explore the context of the banel and the context of the ban. So for this I decided to create a study in which I was going to look at a group of selected bundles at random.

[00:27:35]Okay, randomly selected but they were all non-trigger bands.

[00:27:42]I used all the five word bundles that I identified in the introductions that were nine expressions and I chose 22 four word bundles because they were extremely frequent. They repeated 40 times or more in my corpus that had one million words.

[00:28:07]So first I studied the context meaning the context before and after but looking at the structure to see what type of a structures the bundle was linking.

[00:28:25]After that I did two semantic analysis.

[00:28:29]The first one is a semantic pro and the other one was the semantic preference analysis and for that you have to look at the left code text and the right code text in separate sessions and I'm going to explain to you why. I had a group of PhD students this was um a couple of years ago. These students are all professors at universities now and some of them work at Duolingo.

[00:29:04]So they had they work in the industry and they um we did a group of rating and the analysts so that we could have some interrator reliability.

[00:29:18]H they were called the bundles and beers society because these four students would get together discuss bundles and drink beer. I don't drink alcohol so I would go and drink a diet coke but they would drink uh beers and they would talk about bundles. Well, we met at school, of course, without beers a couple of times, and we did a pilot analysis to see how reliable our analysis was compared, how how they correl we correlated. And for the grammatical analysis, the correlation was very high. It was 0.94 which is something that is not very surprising because out of those the five people three people had studied grammar together so they had the same meta language and then uh for semantic pro studies we did 50% of the occurrences of the bundles and we had an interrator reliability of 08 80%. which for semantic analysis is pretty high too.

[00:30:31]Okay. Um I um so this is what we did. We looked at the bundle and we looked at the language that was to the left of the bundle and then to the language that was to the right of the bundle. And of course depending on the structure of the bundle it was going to be accompanied by different type of structures. So noun phrases with phrasal modification transitive verb phrases punctuation because some bundles start the sentence so they can or the clause so they can come after a period or they can come after a comma or they can come after a colon.

[00:31:16]also some noun phrases with closal modifications, so complex noun phrases and many passive constructions.

[00:31:28]To the right of the bundle, there were many noun phrases with phrasal modification, active transiver phrases, noun phrases with clausal post modification, whole clauses, and coordinated noun phrases. But as you can see, they are listed in um frequency order.

[00:31:53]Well, let me give you one example. So, we have the bundle is based on the Okay, this is a closal bundle. It is a bundle that has a verb phrase, which is a passive verb phrase. But if we look at the code text before and after the bundle, we are going to see that there is a tendency for the left code text.

[00:32:23]You have simple noun phrases mostly the second technique, this awareness, this technique, the argument, this study. So just the the um determiner and the noun sometimes with an adjective but not more than that. On the right context you have a few simple noun phrases but mostly complex noun phrases that had mostly noun complement clauses. So you have something that knowledge that AAA is a legitimate dialect with roots that can be traced back to African languages.

[00:33:10]That is a noun complement clause that is complementing the noun knowledge. So this is a complex noun phrase or the technique is based on the fact that the time period of the analysis is crucial.

[00:33:24]So this is also a noun compliment cause.

[00:33:26]So this is a tendency as you can see you have the the lexical bundle is based on the that is going to have a left cex with a simple noun phrase and a right cex with a complex noun phrase that has a noun which is a shell noun most of the time followed complemented by a noun complement clause.

[00:33:52]This is something that I would have never realized if I hadn't done this study because we have just looked at the bundle with the function but we hadn't looked at the structures around this adds a lot of information for me to share with my students when I teach academic writing because I can now tell them the simple function of the bundle but I can also tell them what is the tendency of the structure structures that function around the bundle.

[00:34:28]Now, semantic proies um the identification of semantic processes is um a procedure that has been done to analyze lexical items. So nobody had used semantic pros or preferences for lexical bundles until probably 2011 or 2013 that I did a study with one of my PhD students Jack Hardy. We compared comparable lexical bundles in Spanish and in English to see if they were used with similar semantic proies and preference across preferences across the two languages.

[00:35:12]So a semantic proy is the way in which the speaker or the writer express the some evaluation some sort of evaluation.

[00:35:24]So it is considered negative or positive but because my corpus is a corpus of research articles in fact research article introductions I try to put a neutral pro too because sometimes in fact in academic writing and in research articles there is a lot of neutral information. The idea is when you read the left co the left context or the right conte context there's going to be some word that is going to tell you whether that discourse is positive or negative. But in academic writing many times you have words that don't provide that proy that they don't tell you if they are positive or negative.

[00:36:16]Left and right processes have to be analyzed separately because you have to try to avoid biased. When you read the left and it's positive, that makes you tend to think that the right one is going to be positive too. So I separated the examples with the left proies and the right proities and I analyze them like that.

[00:36:40]So let me give you some examples. Here you have the left code text of some bundles. The first bundle is the performance of the and you have for example little attention has been paid.

[00:36:55]So this thing of little attention is pro is giving like a negative evaluation.

[00:37:02]You can have a positive proity. For example, here you have the bundle in the presence of and the left cex says the success of the resistant trait depends primarily on the balance between its selective advantage.

[00:37:19]So you have the word success, resistance, primarily balance and advantage that give you the feeling of a positive evaluation. And finally you have the bundle in the context of the and the left code text is neutral because here it says empirical research of form focus instruction has been conducted. There is no adjective or not positive or negative noun. It's just a fact.

[00:37:53]So by default most of the pro studies were neutral.

[00:38:00]70% both left or right. However, some bundles show tendencies for positive plus neutral or negative plus neutral combinations. Okay, this is always a first step that has to be accomp accompanied by the semantic preference of the bundle and we are going to get there in a little bit.

[00:38:27]Let me show you some examples of proies left and right for you to have a better idea. For example, the first bundle is as a result of and in the first example you have positive positive procity. We show that our approach results in an an increased specificity of predicted transcription factor binding sites as a result of a significant reduction of noise. So both are positive. In the next one you have a negative negative.

[00:39:01]Further evidence came from patients who had suffered damage to broka area in the brain as a result of a stroke, an accident or surgery. So those processories are negative.

[00:39:17]Remember that we did this with an interrator reliability. So we know that we all more or less agreed on the process. The second example, the the second group of examples is for the bundle as well as the and here you have neutral positive merins at all use a simulation model to show that the sequent crops as well as the number of different crops can have an important influence on whe populations. So there you have the important influence that is making it positive. And then you have this one that is neutral neutral. This instrumental measures this instrument measures the solar system aridance as well as the spectral solar aridance. That means it's just describing those procedures.

[00:40:10]Now as I said the semantic process is the first step. Then you have to go onto the semantic preference. I need to blow my nose.

[00:40:26]So semantic preference is understood as the semantic field a words collocates predominantly belong to. So looking at the collocations that accompany the bando you can determine the semantic domain of those collocates.

[00:40:48]So it refers to a semantic field that is um accomplished by the collocational environment of the bundle. This also this procedure has been done for lexical items single words but we extended it to the analysis of lexical patterns. So um when we did the study with the rating group um everybody came up with a name for the domain and then we negotiated one name among the five of us because some sometimes we were all meaning the same but we were using different labels.

[00:41:34]So give me let me give you some examples.

[00:41:38]For example, um you have the bundle at the end of the and in research article introductions at the end of the is used on the left to highlight the results of a study or to provide further argumentation and on the right you are talking about discourse or some sort of document. So look at these examples they are all from the corpus. The conclusion for better management strategies will be highlighting at the end of the paper.

[00:42:13]Such enhancements will be discussed further at the end of the paper.

[00:42:19]Finally, the proposed techniques are presented analytically through a case study application at the end of the paper. So you can very well use all this information when teaching how to write research article introductions. Let me give you another example with the bundle has been shown to.

[00:42:43]Usually on the left the writers are referring to an experiment or they are describing a variable and on the right they are indicating significance or influence. Look at these examples. The use of program schemata has been shown to be effective for the optimalization of logic programs or relaxation training also has been shown to improve memory functioning in the elderly or leptin administration has been shown to reduce the fat mass in rodents. So you have these very set uses for these expressions. You are talking about an experiment or certain variable and you are highlighting indicating some influence or significance in its use.

[00:43:39]So what are the pedagogical implications for this? Well, I have now a lot of arguments to add to my academic class, academic writing class. I can teach trigger bundles together with the moves and steps. So when I teach how to organize an introduction and I teach the move and scheme, the move and step scheme that Stone John Swales um created.

[00:44:11]I can also provide a list of bundles that correlate with a particular move and step.

[00:44:20]However, for non-trigger bundles, now I have a lot of arguments that come from the structures that go around that bundle together with the um preference, the procedes and the preferences. I can tell someone, I can tell my students this particular bundle is usually accompanied by these structures, noun phrases, simple noun phrases, uh like in the case of is based on the and complex complex noun phrases on the right and then I can teach them which are the pros. These ones, this bundle is usually used with a right pro proity that is positive and a left pro that is negative or neutral. And then we can go on to the preference and I say well usually on the left code text this bundle is accompanied by this particular semantic domain and on the right this is the semantic domain that follows the bantto.

[00:45:29]So I did this study for bandim research article introductions and together with Elena Cotto who was my student and is now a professor at Iowa State. We did we replicated the methodology with lexical bundles in method sections and we wrote an article Elena is editing a book. So the the chapter is going to come out in her book whenever the book is ready probably last next year.

[00:45:59]But I think that we need to replicate this type of analysis in other registers so that we have a better picture of how the bundles are used and how they interact with their code texts and their contexts. Okay. How to look at the bundle and the structures that exist before and after the bundle and also the pro studies and the preferences.

[00:46:27]And this is um in honor to the bundles and beers group that as I told you they are now all over the world because there are people in other countries and uh they are very successful. uh so it was very very nice to share this experience of analyzing bundles with them when they were students. There you have my email address. So if you have any questions even though now we have enough time for you to ask me questions h if you have any questions or you need to tell me anything you can send me an email and I will be happy to contact you back.

[00:47:21]Does anybody have a question? I don't know if anybody's going to moderate questions. I think I have >> Yes, they do. I was trying to find the camera and and the microphone. Sorry about that.

[00:47:37]>> That's okay.

[00:47:38]Thank you so much, Vivian. was really interesting and I I wish we had had this lecture two years ago because we worked um a lot with Lexico bundles with one of the students here and he would have enjoyed very much um being able to discuss that with you and know all the the steps that you that you followed in order to identify them because they are kind of hidden as you say they are building blocks but we don't pay much attention right when we are writing research pap >> because they are so simple I mean they are not idiomatic they are very transparent so people don't pay attention to them because they are very frequent expressions and everybody uses them but when you are a novice writer and you have to start using them if you really don't know the function that they perform then it's a pity if you don't use them appropriately because it can give your writing a lot of sophistication because you may be using expressions that other people in your academic community are using and so it makes it gives you some some idea of belonging >> to for an academic community. Yeah.

[00:49:00]>> And there are some questions I have some but I will ask the ones that I that we had on the chat. The first >> Okay. Let me see where because I don't have the chat.

[00:49:11]>> Yeah. Yeah. Um I don't know if you can see because it's on the YouTube but I will read them for you.

[00:49:19]>> Will you read me the questions? I'll be happy to answer.

[00:49:22]>> Okay. Okay. So the first one is by Karolina Carvalu and she says that when you introduce the concept of lexical bundles to your students, do they usually recognize or relate to these expressions in their own first language, for example, Portuguese in Brazil?

[00:49:43]And do you think that raising awareness of lexical bundles in the modern uh tongue could make it easier or more natural for learners to understand and apply them in in a foreign language such as English for academic purposes?

[00:50:00]>> Well, it's this is a very very interesting question because don't forget that I teach academic writing in the United States meaning my students come from many different languages.

[00:50:12]probably I have 20 students in a class and not two students have the same first language. I may have a person that speaks Spanish sitting next to a person that speaks Arabic sit sitting next to a person that speaks Korean as the first language. So there's absolutely no way I can teach lexical bundles in their first languages. Okay? places like particularly because I have done studies of lexical bundles in Spanish and in Portuguese and the expressions are very different in their languages in other languages. For for example, I'm going to give you one example of Spanish which and and Portuguese sorry Spanish and English that I'm sure that in Portuguese is more or less the same. Spanish doesn't have premodification of the noun by another noun. So then in Spanish everything is postmodification of the noun with prepositional phrases. So an expression that would be two nouns in English two words in Spanish would be four or five words because you need to post modify the noun. So you can say labor rights and in Spanish it would be.

[00:51:27]>> So you would have five words. The idea is when you when you identify bundles in Spanish you will have many more expressions than in English. Then maybe they have some similarities because if they if you are comparing the same registers. I did a study many years ago in which I looked at lexical bundles in American English from American history and Argentinian Spanish from Argentinian history and the expressions were many of them were very similar in function but structurally they were very different because the languages are different. So first of all when I teach academic writing I don't speak about lexical functions. I just show the expressions and the function but I don't give metal language because my students come from physics and astronomy engineering veterinary and so they don't care what you call those expressions. What they care is what's the function those expressions are performing that can help them communicate a particular intention.

[00:52:43]>> Yeah.

[00:52:44]>> So I try not to fill them up with metal language because probably they are not going to remember that that is called a lexicon. But if I show them a list of expressions that are used for a particular function they may be able to remember that. So I usually teach a class that is genre based in which I teach them the different communicative functions of the introduction or the methodology and they something that I do first is I try to have them identify the communicative functions when they read papers in their discipline because I think that there is always a very big lack of connection between reading and writing that should be together all the time Because if you can identify functions and exponents in the reading that you do, then you can use that as models when you when you write. But I wouldn't I wouldn't be able to to teach um the the lexical bundle in in the first language of my students because as I said before, my students speak languages that I have absolutely no idea what a bundle is. I did a study of lexical bundles in Mandarin Chinese with a student of mine many years ago and we gave up halfway because it was like something so complicated. We had a corpus that was staged for words because you know that in Chinese, Mandarin Chinese, you don't have words, you have um signs and so it was very difficult to know what a lexical bundle was in Chinese. So I rather go on with my language, the languages that I know and I can do Spanish, Portuguese, probably Italian, but then that's it. And and it's I think it's a great question. I understand that if you are teaching academic writing in Brazil where most of your students speak Brazilia, Brazilian Portuguese, you are going to be able to draw some comparisons the same as if I were teaching in Argentina. But I unfortunately I have I don't have that experience. I don't have the experience.

[00:54:57]>> Um and then going a little bit further than like moving from working with future researchers or researchers when they are writing their own papers.

[00:55:07]Have you um had any experience in helping future teachers or English instructors on how to teach lexical bundles because we we have this course here. We have a teacher training course and we would like to make them aware of like tools or ways that they can use in order to prepare their classes and teach like school bundles. Um I teach prospective teachers because our master students most of them are going to be teachers of English as a foreign language or as a second language depending but the bundle that this is something that is very important. The bundle is bound to the register.

[00:55:58]So it will all depend on what your students are going to teach. If they are going to teach academic registers, well, the idea is they are going to have to try to find a group of expressions that are very frequent in that particular register that they are going to teach because I mean unfortunately bundles are not they don't go across registers.

[00:56:20]There are very few expressions that repeat from register to register. So it will have to be um that students are teaching I don't know um argumentative writing because there are bundles that come from argumentative writing. And so then the idea is you teach the expression and the function. So you teach a function and a couple of expressions that are used to convey that function. That's the way. There's nothing you can do with the expression by itself. It has to be tied to the register.

[00:56:55]>> Yeah, I see. I think it's um connected to Sandra's question. Uh she asked if you could uh could we incorporate this knowledge about lexical bundles for essay writing in a general English course as well?

[00:57:13]Yeah, I mean for for essay writing, yes, of course, because there are specific bundles that are used even in the five paragraph essay because you start with a with an introduction, an opening statement, making a claim and then you counter claim and then you um provide evidence and then you conclude and there are expressions that are used for each of those functions. Okay. So there is a um there is an article I think in ESP or the journal of English for academic purposes by Shin Shin N. She in probably 2018 2019 she conducted a study in which she compared argumentative writing written by Korean students writing in English and native speakers writing in English with the same prompts was the same prompts the same conditions and she got a list of bundles. So those are something that can be used to teach argumentative writing but um she didn't find a lot of differences. In fact, because both groups are novice writers and they are struggling with academic writing. So there were like there were more similarities than differences. The problems that they had were that they were using a bundle with um a function that was not really appropriate for a particular intention. But they were everybody was doing the same, you know.

[00:58:49]And of course the native speakers had other issues because they there are other issues that you have when you are writing in English. But um it was not surprising to see that they had many things in common because they were they are both developing as writers both groups.

[00:59:09]>> Yeah, that's very interesting.

[00:59:12]Um Karolina thank you. She said, "Thank you so much, professor." And Sandra has made a comment. Great lecture.

[00:59:22]>> Thank you.

[00:59:23]>> Yeah, I have to say that I have a post-graduate class in the evening and >> usually they are working during the day, but they they have a task. Their assignment is to to watch the lecture >> and then and then I'm gonna make some questions and if they have any comments I I if you allow me I can send to you.

[00:59:47]>> Absolutely. Yes. Yes. Send me any questions that they have. Yeah, >> no problem.

[00:59:53]>> So, thank you so much Vivana for >> Thank you for inviting me >> for being with us. It's really a pleasure and honor and I hope to meet you in person next time.

[01:00:04]>> Yeah, I hope that too because I always wanted to meet with you even when I was in Porto when I was in Sa Paulo in 2023 but we could never meet. So maybe in the >> second semester usually it's very busy but we will manage to meet.

[01:00:21]>> Okay, wonderful. Thank you so much.

[01:00:23]Thank you and I hope you have a a lovely uh evening today.

[01:00:28]>> Thank you so much.

[01:00:30]>> And okay then I thank everyone who attended and the few the ones that will watch. Thank you so much.

[01:00:47]monkey.