Good thing I could recall most of the information about priming thanks to you! I will take my final oral exam next January and I needed a small kickstart from a video like this to start thinking of a possible topic.
Do you see the same cross-lingual priming effects when you translate the second word back to the original language? Can two same-language words be connected through a known second language? To use the "coin" example, does the french "coin" prime French-English bilingual speakers for "argent", or is that too distant a leap?
awesome video, it must take a lot of time to clarify something so complicated. But you had me lost with the example where the priming occurred during "the corner" than the bugs and why spy was dropped.
+walltowall5 Thanks for the question! And we're glad you liked the video. Okay, so it works like this: if we check immediately after someone reads "bugs", even if the context makes it clear we mean the insect kind, your mind still activates "spy". "Ant" might be more relevant, but it doesn't stop us from hauling up "spy" too. But that's only when we check immediately after "bugs". If we wait a couple more words, and check between "the" and "corner" in that sentence, we've already realized that the "spy" meaning is irrelevant, and so we've dropped it. At that point, only "ant" gets the speed bonus, and "spy" is only retrieved as quickly as "sew". Basically, just two short words later, "spy" isn't important anymore, and we've already let it go. It's only immediately after where it makes a difference, and "spy" gets a boost similar to "ant". Does that makes sense? ^_^
+The Ling Space yeah that really helped me understand how each proceeding word affects the connections that are made and let go to make sense of the sentence. thanks a bunch!
Can you guys do an episode on Machine Translation and how language models are constructed? I have a better understanding of it after watching the video but would like to hear your explanation of it. Specially when it comes to how the syntax affect how n-gram Statistical Machine Translation models, since I thought that the SMT takes into account only how frequency the word sequence might appear, not it's correlation with each other.
+Study Room Thanks for the suggestion! We can add it to our topic list, but I think that's a topic we'll have to do a bunch of research on before we can do it justice, so it may be a while before it comes up. But still, it'd be an interesting one! I think I'd like to talk about that, too. We do have a speech recognition one coming up, though, so that should be fun. ^_^
Yeah, and these aren't even the end of the kinds of tests people have devised! As an experimental linguist, I find this sort of stuff really fascinating; I'm looking forward to making more videos about this kind of thing in the future. ^_^
+Soupy Garden path sentences stem from a related coping mechanism for how fast we read and speak, but it's not really the same root of the issue. Instead, we might say that garden path sentences stem from making decisions about sentence structure before we've gotten all of the words we need to be sure. And in fact, we will be saying more of this soon, as we've got a parsing episode coming up in a month or so that'll talk about that. Unrelatedly, I really like your Yousuke icon; he's definitely a personal favourite character of mine. ^_^
6:43 idk whether the sentence you use was part of the original study or not ... but any way, the fact that the word spiders was part of the sentence seems like a confounder to me if you want to check for semantic priming with the target word ''spy''. I know just some seconds later you talk about the decay latency of the priming, but still... Was this intentional ? Nevertheless, Great work ! Thanks for putting all this knowledge out here -
thanks for a cool video!!! I am doing a review of experimtal designs for elicting a N400 ERP with word pairs. But im also curious about priming with pictures and words; do you have any good papers to suggest? thank you!!!
+Soupy Is there any part in particular you mean? There have been studies in Japanese and Chinese as well that show priming with kanji, although not necessarily along as many lines. Japanese is actually particularly interesting, since it uses both ideographic characters (kanji) and syllabic alphabets (hiragana and katakana). It looks like semantic priming is stronger for kanji, but phonological priming stronger with the two kana systems. But we can definitely still say priming is at work in these languages, as well! Check Chen, Yamauchi, Tamaoka and Vaid (2007) for a quick paper looking at some of the differences: people.tamu.edu/~takashi-yamauchi/Reprints/Chen%20(PBR)%202007.pdf. It's far from the only study, but it's a short and easy read. ^_^
I may be wrong, but I think it is essentially building a new thought path through connective and self referential meanings or words and thoughts. This could take time to realise and need reinforcement. Or we could take immediate meaning. Like hearing a song and not thinking much of it, but on 2nd or 3rd listen it connects and has meaning. The first listen primes you for it. It creates a new path in your brain that you can see before but won't follow at first. The more priming that exists, the more likely you are to follow the path and find the meaning. That is how I view it. But i could also have been primed to misunderstand it.
+Pakanahymni We talked with a couple of sound engineer people at VidCon and described our setup to them, and they suggested some changes we could make in terms of gain control through the system that they thought were the problem. And I think it solved a lot of it! We were really happy to find some help, and I'm glad you noticed. ^_^
+vc07441 Yeah, these networks definitely make it easier to pull up words and build connections. It's also part of why reading unfamiliar material is harder - you don't have the ties to yank up the words you need, and so you need to build them up.
+lolidays27 Maybe! It depends where you're saying you don't have much knowledge. Is it for the programming part, or for the linguistic part? And how much time do you have to get it ready? ^_^
+lolidays27 Hmmm. Well, if you're willing to play around with it, it might be easier than you think! There's often a bunch of documentation around for these programs, and if you don't want to do anything complex, it's not hard to get started. If you want to do priming, you could use DMDX - it's free, and there are some good tutorials and walkthroughs online, like here: www.mrc-cbu.cam.ac.uk/people/matt.davis/dmdx-tutorial/
+The Ling Space Hey, I know it's been a long time, but I just wanted to thank you for your help :) I just finished testing it out on people, though the difference between reaction times was a bit less extreme than I thought (34 milliseconds) :P
+Soupy Well, as we talk about later in the episode, there are masked priming studies (some links in the description) that show that priming effects, both phonological and semantic, can be found at 50 milliseconds, which is 1/20 of a second. If you multiply that out, that'd be 1200 words a minute. So I think it's probably the case that even for speed reading, there's no reason to think it's fundamentally different! ^_^
+Corey Levinson Priming studies aren't super hard to put together, but it does take a bit of planning! But there are some free software things you could play around with that linguists use to design these sorts of studies. Although for priming, I'd say that some of the software that gets used, like Matlab or ePrime, does cost money, there are other resources. There could be ways for you to involved in research otherwise, too, though. ^_^
1:50 "And then as soon as you identify the one you actually need, you just drop all of the rest in a matter of a tenth of a second or two." -- Were these studies done with neurotypical people? Because in my experience as a neurodivergent person (Tourette syndrome, ADHD), I often make the wrong interpretation and make a fool of myself. With your "bug" example, for instance, my interpretation as "error in computer code" persisted for a consciously measurable span of time even after the clue of "in the corner". I mean, someone can have their computer set up in the corner of a room.
This video was excellent in explaining and so helpful for studying for my psychology exam on Friday. Thanks so much!
Good thing I could recall most of the information about priming thanks to you! I will take my final oral exam next January and I needed a small kickstart from a video like this to start thinking of a possible topic.
Do you see the same cross-lingual priming effects when you translate the second word back to the original language? Can two same-language words be connected through a known second language? To use the "coin" example, does the french "coin" prime French-English bilingual speakers for "argent", or is that too distant a leap?
seriously you guys deserve more subs , thank you for everything
Excellent explanation! I was worried about understanding some psycholinguistics concepts, but you made me get there! Thx!
awesome video, it must take a lot of time to clarify something so complicated. But you had me lost with the example where the priming occurred during "the corner" than the bugs and why spy was dropped.
+walltowall5 Thanks for the question! And we're glad you liked the video. Okay, so it works like this: if we check immediately after someone reads "bugs", even if the context makes it clear we mean the insect kind, your mind still activates "spy". "Ant" might be more relevant, but it doesn't stop us from hauling up "spy" too. But that's only when we check immediately after "bugs". If we wait a couple more words, and check between "the" and "corner" in that sentence, we've already realized that the "spy" meaning is irrelevant, and so we've dropped it. At that point, only "ant" gets the speed bonus, and "spy" is only retrieved as quickly as "sew". Basically, just two short words later, "spy" isn't important anymore, and we've already let it go. It's only immediately after where it makes a difference, and "spy" gets a boost similar to "ant". Does that makes sense? ^_^
+The Ling Space yeah that really helped me understand how each proceeding word affects the connections that are made and let go to make sense of the sentence.
thanks a bunch!
Can you guys do an episode on Machine Translation and how language models are constructed? I have a better understanding of it after watching the video but would like to hear your explanation of it.
Specially when it comes to how the syntax affect how n-gram Statistical Machine Translation models, since I thought that the SMT takes into account only how frequency the word sequence might appear, not it's correlation with each other.
+Study Room Thanks for the suggestion! We can add it to our topic list, but I think that's a topic we'll have to do a bunch of research on before we can do it justice, so it may be a while before it comes up. But still, it'd be an interesting one! I think I'd like to talk about that, too. We do have a speech recognition one coming up, though, so that should be fun. ^_^
Is this the same as "the lexical decision task"?
It is interesting to me how you can construct experiments to test certain theories of how the brain works.
Yeah, and these aren't even the end of the kinds of tests people have devised! As an experimental linguist, I find this sort of stuff really fascinating; I'm looking forward to making more videos about this kind of thing in the future. ^_^
+Elia Magyar Cool! Thanks for passing that along. I didn't know! ^_^
Incredible. Thank you for that. You can see he didn't err about in his choice of content. Smarty pants.
Your videos are excellent!! They have helped me a lot!
Sorry for asking so many questions but, does this relate to garden path sentences?
+Soupy Garden path sentences stem from a related coping mechanism for how fast we read and speak, but it's not really the same root of the issue. Instead, we might say that garden path sentences stem from making decisions about sentence structure before we've gotten all of the words we need to be sure. And in fact, we will be saying more of this soon, as we've got a parsing episode coming up in a month or so that'll talk about that.
Unrelatedly, I really like your Yousuke icon; he's definitely a personal favourite character of mine. ^_^
+The Ling Space thanks for answering all my questions!
on a now related note, tennis for the compliment. :D
6:43 idk whether the sentence you use was part of the original study or not ... but any way, the fact that the word spiders was part of the sentence seems like a confounder to me if you want to check for semantic priming with the target word ''spy''. I know just some seconds later you talk about the decay latency of the priming, but still...
Was this intentional ?
Nevertheless,
Great work ! Thanks for putting all this knowledge out here -
you are super smart. period.
+CuteLittleAnimals Haha, thanks! And thanks for watching. ^_^
Excellent
Great video - thanks!
i love this guy
also how are you that much better than my linguistics courses lol
Great video
Can you please explain masked priming?
thanks for a cool video!!! I am doing a review of experimtal designs for elicting a N400 ERP with word pairs. But im also curious about priming with pictures and words; do you have any good papers to suggest? thank you!!!
Amazing!!!
7:57 Was this only done with languages that use alphabets?
+Soupy Is there any part in particular you mean? There have been studies in Japanese and Chinese as well that show priming with kanji, although not necessarily along as many lines. Japanese is actually particularly interesting, since it uses both ideographic characters (kanji) and syllabic alphabets (hiragana and katakana). It looks like semantic priming is stronger for kanji, but phonological priming stronger with the two kana systems. But we can definitely still say priming is at work in these languages, as well! Check Chen, Yamauchi, Tamaoka and Vaid (2007) for a quick paper looking at some of the differences: people.tamu.edu/~takashi-yamauchi/Reprints/Chen%20(PBR)%202007.pdf. It's far from the only study, but it's a short and easy read. ^_^
Thanks. Could you tell me what you mean by "priming effect" ?
I may be wrong, but I think it is essentially building a new thought path through connective and self referential meanings or words and thoughts. This could take time to realise and need reinforcement. Or we could take immediate meaning. Like hearing a song and not thinking much of it, but on 2nd or 3rd listen it connects and has meaning. The first listen primes you for it. It creates a new path in your brain that you can see before but won't follow at first. The more priming that exists, the more likely you are to follow the path and find the meaning.
That is how I view it. But i could also have been primed to misunderstand it.
Hey the audio sounds good, the ping-pong echo is gone for the most part! What did you guys change?
+Pakanahymni We talked with a couple of sound engineer people at VidCon and described our setup to them, and they suggested some changes we could make in terms of gain control through the system that they thought were the problem. And I think it solved a lot of it! We were really happy to find some help, and I'm glad you noticed. ^_^
The Ling Space Great! Now my language apparatus has to work less to decode your messages!
What was the word that was flashed at 7:45? I'm having a hard time freezing on it. Logan? Login?
Yep, it's Logan. It does go by pretty fast, but that was the point. ^_^
oh this is why we read faster when we read more!
+vc07441 Yeah, these networks definitely make it easier to pull up words and build connections. It's also part of why reading unfamiliar material is harder - you don't have the ties to yank up the words you need, and so you need to build them up.
true.really appreciate what you are doing!please keep it up!!
I'm thinking of doing something linked to this for my science fair. Do you think I could set up a program to test it without much knowledge? :/
+lolidays27 Maybe! It depends where you're saying you don't have much knowledge. Is it for the programming part, or for the linguistic part? And how much time do you have to get it ready? ^_^
+The Ling Space It's the programming part :/
I'd have a bit more than two months to get ready
+lolidays27 Hmmm. Well, if you're willing to play around with it, it might be easier than you think! There's often a bunch of documentation around for these programs, and if you don't want to do anything complex, it's not hard to get started. If you want to do priming, you could use DMDX - it's free, and there are some good tutorials and walkthroughs online, like here: www.mrc-cbu.cam.ac.uk/people/matt.davis/dmdx-tutorial/
+The Ling Space thank you so much! :D
+The Ling Space Hey, I know it's been a long time, but I just wanted to thank you for your help :) I just finished testing it out on people, though the difference between reaction times was a bit less extreme than I thought (34 milliseconds) :P
I wonder if this is related to Warf/Sapir.
I saw the word it was Logan
dude your videos are cool
Thanks! Glad you like them! ^_^
Is the shirt a reference to "Veronica Mars"? That's all I can think about through this video. 😂🙈
+boriquabbbls I think so, it ties into the theme of picking up and putting together clues. Also the first target word being 'sheriff' is a fun tie-in.
+boriquabbbls Yep, it is a Veronica Mars related video! I really like that shirt, too. We like it when people catch our themes for the week. ^_^
+boriquabbbls haha, same! as soon as i realized i couldn't stop thinking about it
0:53 have there been many studies on even faster wpm? I know squirt.io and sprtiz allow you to get up to ~750+ wpm with practice.
+Soupy Well, as we talk about later in the episode, there are masked priming studies (some links in the description) that show that priming effects, both phonological and semantic, can be found at 50 milliseconds, which is 1/20 of a second. If you multiply that out, that'd be 1200 words a minute. So I think it's probably the case that even for speed reading, there's no reason to think it's fundamentally different! ^_^
wow, i really want to try doing the experiments you talked about with some bilingual friends !
+Corey Levinson Priming studies aren't super hard to put together, but it does take a bit of planning! But there are some free software things you could play around with that linguists use to design these sorts of studies. Although for priming, I'd say that some of the software that gets used, like Matlab or ePrime, does cost money, there are other resources. There could be ways for you to involved in research otherwise, too, though. ^_^
Im using this topic in my thesis
But i can't find the suitable research question for this topic
Can anyone help me 🙏😅
surely just look at the links provided, post them into google scholar and check the box "cited by", with the date range set to 2015+?
loud "-elow" as you look down at your mic... 😀
+Inez Allen Yeah, that was a more extreme look down than usual. I'll try to be more careful! ^_^
it wasn't technically //bad// just funny.
LOGAN!
1:50 "And then as soon as you identify the one you actually need, you just drop all of the rest in a matter of a tenth of a second or two." -- Were these studies done with neurotypical people? Because in my experience as a neurodivergent person (Tourette syndrome, ADHD), I often make the wrong interpretation and make a fool of myself. With your "bug" example, for instance, my interpretation as "error in computer code" persisted for a consciously measurable span of time even after the clue of "in the corner". I mean, someone can have their computer set up in the corner of a room.
Excellent