1 00:00:00,000 --> 00:00:02,660 Welcome to the deep dive. Today, we're diving 2 00:00:02,660 --> 00:00:05,679 into something huge, really fundamental. It's 3 00:00:05,679 --> 00:00:08,800 about where are our most complex ideas? Think 4 00:00:08,800 --> 00:00:12,320 about things like infinity or evolution or even 5 00:00:12,320 --> 00:00:15,439 just the concept of seven. Where do they actually 6 00:00:15,439 --> 00:00:17,579 come from? It's a massive question. Right. Are 7 00:00:17,579 --> 00:00:20,699 they built up bit by bit from what we see and 8 00:00:20,699 --> 00:00:24,440 hear? Or is the mind sort of preloaded? ready 9 00:00:24,440 --> 00:00:26,160 for them. Well, that's exactly what we're digging 10 00:00:26,160 --> 00:00:28,100 into today. We're moving past the old armchair 11 00:00:28,100 --> 00:00:30,399 philosophy, looking at the actual science, the 12 00:00:30,399 --> 00:00:32,640 cognitive development research. And what it strongly 13 00:00:32,640 --> 00:00:34,640 suggests, especially the work on concept origins 14 00:00:34,640 --> 00:00:37,500 we're discussing, is that the blank slate idea, 15 00:00:37,759 --> 00:00:40,579 it just doesn't hold up. OK. So our mission, 16 00:00:40,899 --> 00:00:43,479 really, is to get clear on this relationship 17 00:00:43,479 --> 00:00:45,539 between our built -in mental structures, this 18 00:00:45,539 --> 00:00:48,259 core cognition, and how we get to mature abstract 19 00:00:48,259 --> 00:00:50,520 thought, and critically, how language fits in. 20 00:00:50,649 --> 00:00:52,929 Because it seems language isn't just for communication, 21 00:00:53,250 --> 00:00:55,090 it's somehow crucial for building these more 22 00:00:55,090 --> 00:00:57,729 complex concepts. Language is a tool for thought 23 00:00:57,729 --> 00:00:59,770 itself. Exactly. We want to figure out what the 24 00:00:59,770 --> 00:01:02,549 mind starts with, what its initial limits are, 25 00:01:02,570 --> 00:01:05,650 and then how language helps it leap over those 26 00:01:05,650 --> 00:01:07,209 limits. OK, so let's map this out. We'll start 27 00:01:07,209 --> 00:01:08,989 with the basics. What is this core cognition? 28 00:01:09,489 --> 00:01:12,310 Then we'll look at how we handle quantity numbers 29 00:01:12,310 --> 00:01:14,590 before language comes into the picture. They're 30 00:01:14,590 --> 00:01:17,950 these two fascinating systems. Powerful but limited. 31 00:01:18,049 --> 00:01:21,409 Right. And then the main event. Yeah. How language 32 00:01:21,409 --> 00:01:24,569 acts is this. This bootstrapper, creating new 33 00:01:24,569 --> 00:01:26,569 ways of thinking. And we'll use learning to count, 34 00:01:27,030 --> 00:01:29,409 learning about natural numbers as our big case 35 00:01:29,409 --> 00:01:32,049 study. That's the plan. It really showcases the 36 00:01:32,049 --> 00:01:34,430 process. All right. So for a long, long time, 37 00:01:34,590 --> 00:01:37,430 the dominant idea, you know, Locke, Hume, the 38 00:01:37,430 --> 00:01:40,349 empiricists, was that we start with just basic 39 00:01:40,349 --> 00:01:42,810 sensations, a flash of red, a loud noise, and 40 00:01:42,810 --> 00:01:44,510 everything else. All our complex ideas are just 41 00:01:44,510 --> 00:01:46,780 combinations of these. But this core cognition 42 00:01:46,780 --> 00:01:49,540 idea challenges that head on. It really does. 43 00:01:49,920 --> 00:01:52,280 The claim isn't that the mind is empty waiting 44 00:01:52,280 --> 00:01:55,280 for data. It's that the mind starts with these, 45 00:01:55,280 --> 00:01:58,620 well, highly structured, innate mechanisms. Think 46 00:01:58,620 --> 00:02:01,019 of them like specialized apps pre -installed 47 00:02:01,019 --> 00:02:03,819 on a computer. Apps for what, specifically? Apps 48 00:02:03,819 --> 00:02:06,140 designed to build representations about specific 49 00:02:06,140 --> 00:02:08,919 things in the world. Core domains like objects, 50 00:02:09,020 --> 00:02:11,280 knowing what a physical thing is and how it behaves. 51 00:02:11,939 --> 00:02:14,139 Or agents understanding things that act with 52 00:02:14,139 --> 00:02:16,939 intention. and even basic number or quantity. 53 00:02:17,500 --> 00:02:19,280 They're essentially domain -specific learning 54 00:02:19,280 --> 00:02:23,039 devices. So not just raw processing power, but 55 00:02:23,039 --> 00:02:25,939 tools tailored for specific jobs. Precisely. 56 00:02:26,159 --> 00:02:28,340 They're evolutionary solutions. The world reliably 57 00:02:28,340 --> 00:02:31,280 contains objects and agents, so evolution gave 58 00:02:31,280 --> 00:02:33,439 us a head start in understanding them. And this 59 00:02:33,439 --> 00:02:35,460 isn't just about simple reflexes, right? We're 60 00:02:35,460 --> 00:02:39,319 talking actual concepts. Yes, that's key. These 61 00:02:39,319 --> 00:02:41,979 systems produce representations with specific 62 00:02:41,979 --> 00:02:44,430 content. Take the indigo buntings example from 63 00:02:44,430 --> 00:02:46,830 the source material. It's a fantastic illustration 64 00:02:46,830 --> 00:02:49,430 outside of humans. Ah yes, the songbirds that 65 00:02:49,430 --> 00:02:51,729 navigate by the stars. Tell us about that. Okay, 66 00:02:52,289 --> 00:02:55,030 so these tiny birds migrate thousands of miles. 67 00:02:55,210 --> 00:02:58,330 How? They use the stars. But evolution couldn't 68 00:02:58,330 --> 00:03:00,949 just hardwire a star map because the sky changes 69 00:03:00,949 --> 00:03:03,030 over time, right? Right, precession and all that. 70 00:03:03,349 --> 00:03:06,650 Exactly. So instead, evolution gave them an innate 71 00:03:06,650 --> 00:03:09,780 computational mechanism. It's specifically designed 72 00:03:09,780 --> 00:03:12,039 to look at the pattern of the night sky's rotation 73 00:03:12,039 --> 00:03:14,659 and identify the center point of that rotation. 74 00:03:14,659 --> 00:03:16,840 And the center point is? North or very close 75 00:03:16,840 --> 00:03:19,139 to it. So the bunting doesn't just learn a pattern 76 00:03:19,139 --> 00:03:21,780 of lights. It computes and represents north an 77 00:03:21,780 --> 00:03:24,879 abstract direction. It's a conceptual representation 78 00:03:24,879 --> 00:03:28,520 generated by a mechanism tailored just for celestial 79 00:03:28,520 --> 00:03:31,139 navigation. That's core cognition in action. 80 00:03:31,539 --> 00:03:34,139 Wow. Okay, so it's a specific computation leading 81 00:03:34,139 --> 00:03:36,620 to a concept, but How is that different from 82 00:03:36,620 --> 00:03:39,360 the concepts we adults have? If the bunting has 83 00:03:39,360 --> 00:03:41,979 a concept of north and an infant has a concept 84 00:03:41,979 --> 00:03:45,300 of object, how does that differ from, say, my 85 00:03:45,300 --> 00:03:47,800 concept of democracy or calculus? The crucial 86 00:03:47,800 --> 00:03:50,240 difference seems to be the format of the representation 87 00:03:50,240 --> 00:03:53,500 and its expressive power. Core cognition representations 88 00:03:53,500 --> 00:03:56,539 are thought to be largely iconic or analog. Like 89 00:03:56,539 --> 00:03:59,979 a picture? Sort of, yes. Or like a mental model 90 00:03:59,979 --> 00:04:02,629 that resembles what it represents. Think of the 91 00:04:02,629 --> 00:04:05,069 bunting's north representation, maybe like an 92 00:04:05,069 --> 00:04:07,689 internal compass needle, or the infant's object 93 00:04:07,689 --> 00:04:09,949 representation like a little mental simulation 94 00:04:09,949 --> 00:04:12,569 of the object's trajectory. It's tied closely 95 00:04:12,569 --> 00:04:15,909 to perception and action. Mature concepts, though, 96 00:04:16,089 --> 00:04:18,829 the ones we build later, often rely on discrete 97 00:04:18,829 --> 00:04:21,769 arbitrary symbols. Like words, or mathematical 98 00:04:21,769 --> 00:04:24,430 symbols. Exactly, language -like symbols. These 99 00:04:24,430 --> 00:04:26,670 symbols aren't constrained by resemblance. The 100 00:04:26,670 --> 00:04:29,670 word dog doesn't look like a dog. its power comes 101 00:04:29,670 --> 00:04:31,810 from its relationship to other symbols in a system, 102 00:04:31,990 --> 00:04:35,110 the rules of grammar, logic, mathematics. So 103 00:04:35,110 --> 00:04:37,470 you can have a picture of round, maybe, but not 104 00:04:37,470 --> 00:04:41,730 a picture of electron or the number 3 ,462 ,179. 105 00:04:41,889 --> 00:04:44,769 For those, you need symbols. Precisely. Evolution 106 00:04:44,769 --> 00:04:47,470 equipped us to perceive roundness directly. But 107 00:04:47,470 --> 00:04:49,709 concepts like electron or large exact numbers 108 00:04:49,709 --> 00:04:51,790 require these symbolic systems, built through 109 00:04:51,790 --> 00:04:53,910 learning, social interaction, complex reasoning. 110 00:04:54,439 --> 00:04:56,300 things that fix what those symbols refer to. 111 00:04:56,500 --> 00:04:58,439 So core systems are conceptual, but limited. 112 00:04:58,720 --> 00:05:00,740 The bunting nose north, but it doesn't have a 113 00:05:00,740 --> 00:05:02,779 theory of astrophysics, you know? Right, it's 114 00:05:02,779 --> 00:05:06,259 a practical, constrained concept. Okay, this 115 00:05:06,259 --> 00:05:08,180 difference in format really hits home when we 116 00:05:08,180 --> 00:05:11,319 look at the object concept again. Piaget thought 117 00:05:11,319 --> 00:05:15,149 babies built this idea slowly. But the core cognition 118 00:05:15,149 --> 00:05:17,509 view says... It says Piaget was mistaken about 119 00:05:17,509 --> 00:05:21,009 the timeline. Modern research using clever methods 120 00:05:21,009 --> 00:05:23,550 shows that even very young infants, we're talking 121 00:05:23,550 --> 00:05:26,290 two to five months old, already represent objects 122 00:05:26,290 --> 00:05:29,430 as solid continuous things that exist even when 123 00:05:29,430 --> 00:05:32,170 out of sight. Object permanence way earlier. 124 00:05:32,329 --> 00:05:34,810 Way earlier. And they understand basic physical 125 00:05:34,810 --> 00:05:37,069 rules like solidity two objects can't be in the 126 00:05:37,069 --> 00:05:39,389 same place at the same time. How do we know that? 127 00:05:39,449 --> 00:05:41,509 The rotating screen experiment is the classic, 128 00:05:41,709 --> 00:05:43,750 isn't it? It is. It's elegant. Basically, you 129 00:05:43,750 --> 00:05:45,810 show an infant a screen rotating back and forth 130 00:05:45,810 --> 00:05:47,670 like a drawbridge. Simple enough. They get used 131 00:05:47,670 --> 00:05:50,810 to it. OK. Then you place a solid block behind 132 00:05:50,810 --> 00:05:53,350 the screen where the screen should stop. In the 133 00:05:53,350 --> 00:05:55,550 possible event, the screen rotates and stops 134 00:05:55,550 --> 00:05:57,310 when it hits the hidden block. Makes sense? Uh 135 00:05:57,310 --> 00:05:59,829 -huh. In the impossible event, the screen appears 136 00:05:59,829 --> 00:06:02,069 to rotate right through the space where the block 137 00:06:02,069 --> 00:06:05,149 is hidden. And infants look reliably longer at 138 00:06:05,149 --> 00:06:07,939 that impossible event. Why? Because it violates 139 00:06:07,939 --> 00:06:11,019 their expectations. Exactly. To be surprised, 140 00:06:11,420 --> 00:06:13,680 they must have represented the block continuing 141 00:06:13,680 --> 00:06:16,120 to exist behind the screen, and they must have 142 00:06:16,120 --> 00:06:19,000 expected it to stop the screen's motion because 143 00:06:19,000 --> 00:06:21,899 of solidity. This understanding is present so 144 00:06:21,899 --> 00:06:24,639 early, it has to be part of that initial core 145 00:06:24,639 --> 00:06:27,339 cognition toolkit. It fundamentally challenges 146 00:06:27,339 --> 00:06:29,899 the idea that object understanding is slowly 147 00:06:29,899 --> 00:06:32,480 constructed or purely dependent on language. 148 00:06:32,819 --> 00:06:35,569 So... When older babies fail tasks like the A 149 00:06:35,569 --> 00:06:37,670 not B error, where they reach for a toy where 150 00:06:37,670 --> 00:06:40,089 it used to be hidden, that's not because they 151 00:06:40,089 --> 00:06:42,949 lack the object concept. It seems not. It's more 152 00:06:42,949 --> 00:06:44,990 likely an issue with immature executive functions, 153 00:06:45,569 --> 00:06:47,889 difficulty inhibiting a prior action, or managing 154 00:06:47,889 --> 00:06:50,910 complex search plans. The concept is there. The 155 00:06:50,910 --> 00:06:53,189 performance is the problem. Making that distinction 156 00:06:53,189 --> 00:06:55,910 is absolutely vital. The foundation for understanding 157 00:06:55,910 --> 00:06:58,509 physical objects seems to be innate. Okay, objects 158 00:06:58,509 --> 00:07:00,990 are innate. Now let's pivot to quantity, to number. 159 00:07:01,470 --> 00:07:03,129 Most people probably think number comes from 160 00:07:03,129 --> 00:07:05,629 learning to count, right? One, two, three. But 161 00:07:05,629 --> 00:07:07,689 the research shows we have ways of dealing with 162 00:07:07,689 --> 00:07:10,990 quantity before we learn number words. Two distinct 163 00:07:10,990 --> 00:07:13,810 systems, actually. That's right. And understanding 164 00:07:13,810 --> 00:07:16,209 these two pre -linguistic systems is crucial 165 00:07:16,209 --> 00:07:19,050 because their limitations are precisely what 166 00:07:19,050 --> 00:07:21,149 creates the need for language to step in later. 167 00:07:21,290 --> 00:07:22,910 Let's break them down. What's the first one? 168 00:07:23,269 --> 00:07:25,589 The first is often called the parallel individuation 169 00:07:25,589 --> 00:07:28,029 system or the object file system. Its job is 170 00:07:28,029 --> 00:07:30,629 basically to track individual things. Like keeping 171 00:07:30,629 --> 00:07:33,649 tabs on specific objects? Exactly. It creates 172 00:07:33,649 --> 00:07:37,149 a kind of temporary mental workspace with one 173 00:07:37,149 --> 00:07:39,829 file or token for each distinct individual you're 174 00:07:39,829 --> 00:07:41,879 paying attention to. Think of it like having 175 00:07:41,879 --> 00:07:44,680 mental pointers aimed at specific things. It's 176 00:07:44,680 --> 00:07:47,339 great for tracking who did what to whom or where 177 00:07:47,339 --> 00:07:49,420 things are located. Sounds pretty useful for 178 00:07:49,420 --> 00:07:51,879 everyday stuff. But you mentioned a catch, a 179 00:07:51,879 --> 00:07:54,680 limitation. A very sharp limitation. This system 180 00:07:54,680 --> 00:07:57,459 has a hard set size limit. It can typically only 181 00:07:57,459 --> 00:07:59,899 handle about three or four individuals simultaneously. 182 00:08:00,160 --> 00:08:03,579 Only three or four? That seems long. It is. That's 183 00:08:03,579 --> 00:08:05,560 why juggling more than a few items in your head 184 00:08:05,560 --> 00:08:09,399 gets difficult fast. Try tracking five unique 185 00:08:09,399 --> 00:08:12,389 moving dots. It's almost impossible without grouping 186 00:08:12,389 --> 00:08:15,500 them or using some other strategy. The evidence 187 00:08:15,500 --> 00:08:18,379 for this limit in infants is striking, especially 188 00:08:18,379 --> 00:08:20,860 for manual search tasks. OK. Walk us through 189 00:08:20,860 --> 00:08:22,639 one of those experiments. The cracker one is 190 00:08:22,639 --> 00:08:24,800 memorable. Right. Feigenstein and Carey's work. 191 00:08:25,399 --> 00:08:28,439 Imagine a 12 -month -old watching you place crackers 192 00:08:28,439 --> 00:08:32,580 one by one into an opaque box. They love crackers, 193 00:08:32,620 --> 00:08:34,580 so they're motivated to get them. Naturally. 194 00:08:34,720 --> 00:08:37,139 If you put, say, three crackers in the box and 195 00:08:37,139 --> 00:08:39,159 then let the infant reach in and pull out only 196 00:08:39,159 --> 00:08:41,019 two, what do they do? They keep searching for 197 00:08:41,019 --> 00:08:43,419 the third one? Precisely. They search longer. 198 00:08:43,720 --> 00:08:46,259 They seem surprised or frustrated. Their object 199 00:08:46,259 --> 00:08:48,539 file system tracked cracker one, cracker two, 200 00:08:48,600 --> 00:08:51,240 cracker three, saw two removed, and knows one 201 00:08:51,240 --> 00:08:53,700 file is still open. They expect exactly one more. 202 00:08:53,899 --> 00:08:56,019 OK, makes sense. Now for the surprising part. 203 00:08:56,120 --> 00:08:58,860 Here it is. If you put four crackers in the box 204 00:08:58,860 --> 00:09:02,070 and let the infant retrieve only one, they don't 205 00:09:02,070 --> 00:09:04,629 search longer. They act as if the box is empty. 206 00:09:04,759 --> 00:09:06,500 Wait, even though there are still three crackers 207 00:09:06,500 --> 00:09:08,460 left, that's numerically a much bigger difference. 208 00:09:08,539 --> 00:09:10,440 Four versus one is easier than three versus two, 209 00:09:10,539 --> 00:09:13,100 isn't it? Numerically, yes. The 4 .1 ratio is 210 00:09:13,100 --> 00:09:15,440 much easier than 3 .2. But the instant fails. 211 00:09:15,559 --> 00:09:18,059 Why? Because the parallel individuation system 212 00:09:18,059 --> 00:09:20,340 hit its limit. Once you go past three or four, 213 00:09:20,360 --> 00:09:22,799 the system just collapses. It can't maintain 214 00:09:22,799 --> 00:09:25,100 distinct files for each item anymore. It's not 215 00:09:25,100 --> 00:09:27,419 about the arithmetic ratio. It's about the absolute 216 00:09:27,419 --> 00:09:29,440 number of individuals it can track. It proves 217 00:09:29,440 --> 00:09:31,919 this system is about tokens, not abstract values. 218 00:09:32,559 --> 00:09:35,159 Wow, okay, that's a really fun - limit on our 219 00:09:35,159 --> 00:09:38,279 innate ability to handle exact numbers. System 220 00:09:38,279 --> 00:09:41,740 one handles small exact sets up to about three, 221 00:09:41,879 --> 00:09:43,980 but obviously we can deal with larger quantities. 222 00:09:44,279 --> 00:09:46,220 How do we do that without language? How do we 223 00:09:46,220 --> 00:09:48,059 know a crowd of 50 is smaller than a crowd of 224 00:09:48,059 --> 00:09:51,940 100? That brings us to system two, analog magnitude 225 00:09:51,940 --> 00:09:55,000 representations. This system is almost the opposite 226 00:09:55,000 --> 00:09:56,899 of the first one. It's not for tracking individuals, 227 00:09:56,960 --> 00:09:58,779 it's for representing approximate quantities 228 00:09:58,779 --> 00:10:03,269 or numerosity. estimation. Exactly. Think of 229 00:10:03,269 --> 00:10:05,309 it like a mental number line, but it's fuzzy. 230 00:10:05,750 --> 00:10:08,470 It can represent a lot versus a little or more 231 00:10:08,470 --> 00:10:11,669 versus less, but its precision is limited, especially 232 00:10:11,669 --> 00:10:14,470 as numbers get bigger. Its signature constraint 233 00:10:14,470 --> 00:10:17,049 is scalar variability, which follows Weber's 234 00:10:17,049 --> 00:10:19,309 law. Wimber's Law, that sounds technical, break 235 00:10:19,309 --> 00:10:22,009 it down for us. It basically means your ability 236 00:10:22,009 --> 00:10:24,470 to tell two quantities apart depends on their 237 00:10:24,470 --> 00:10:27,049 ratio, not the absolute difference between them. 238 00:10:27,629 --> 00:10:30,250 So distinguishing 10 dots from 20 dots, a 1 .2 239 00:10:30,250 --> 00:10:33,570 ratio is relatively easy, but distinguishing 240 00:10:33,570 --> 00:10:36,990 110 dots from 120 dots is much harder. even though 241 00:10:36,990 --> 00:10:39,049 the absolute difference is still 10, because 242 00:10:39,049 --> 00:10:43,330 the ratio 11 .12 is much closer to 1 .1. The 243 00:10:43,330 --> 00:10:46,129 fuzziness of the representation increases proportionally 244 00:10:46,129 --> 00:10:48,590 with the magnitude. So the bigger the numbers, 245 00:10:48,690 --> 00:10:50,769 the harder it is to tell them apart, unless the 246 00:10:50,769 --> 00:10:52,529 difference is also proportionally bigger. You 247 00:10:52,529 --> 00:10:54,509 got it. You can easily see the difference between 248 00:10:54,509 --> 00:10:57,210 five sheep and 10 sheep. But telling 100 sheep 249 00:10:57,210 --> 00:10:59,830 from 105 sheep just by looking, very difficult. 250 00:11:00,230 --> 00:11:02,190 Your analog magnitude system gives you a rough 251 00:11:02,190 --> 00:11:04,669 sense, but not exact precision. And we see this 252 00:11:04,669 --> 00:11:07,860 system in animals too, right? Absolutely. Classic 253 00:11:07,860 --> 00:11:10,039 studies, like with rats discriminating sequences 254 00:11:10,039 --> 00:11:12,440 of tones or light, say, learning to press a lever 255 00:11:12,440 --> 00:11:15,019 after two flashes, but not after eight. They 256 00:11:15,019 --> 00:11:17,340 can do it, but their accuracy follows Weber's 257 00:11:17,340 --> 00:11:19,860 law perfectly. Their timing or counting gets 258 00:11:19,860 --> 00:11:22,720 proportionally sloppier for larger numbers. It 259 00:11:22,720 --> 00:11:25,720 shows this is a very old shared system for approximate 260 00:11:25,720 --> 00:11:29,460 quantity. OK, so let's recap. We have system 261 00:11:29,460 --> 00:11:31,639 one, exact tracking, but only up to three or 262 00:11:31,639 --> 00:11:34,960 four items, and system two, approximate estimation 263 00:11:34,960 --> 00:11:37,899 for larger quantities. limited by ratio. Neither 264 00:11:37,899 --> 00:11:40,399 of these, on their own or together, can give 265 00:11:40,399 --> 00:11:43,460 us the concept of, say, exactly seven, or the 266 00:11:43,460 --> 00:11:45,299 idea that numbers go on forever, the successor 267 00:11:45,299 --> 00:11:48,870 function. That's the crucial gap. Neither system 268 00:11:48,870 --> 00:11:50,929 can represent the precise cardinal value of sets 269 00:11:50,929 --> 00:11:53,269 beyond three or four. Neither system understands 270 00:11:53,269 --> 00:11:55,149 that for every number n, there's always a next 271 00:11:55,149 --> 00:11:57,850 number n plus one. There's no concept of infinity, 272 00:11:58,090 --> 00:12:01,090 no basis for formal arithmetic. And that limitation, 273 00:12:01,169 --> 00:12:03,269 that's where language becomes essential. It's 274 00:12:03,269 --> 00:12:05,730 the cognitive bottleneck. To get beyond these 275 00:12:05,730 --> 00:12:08,049 core systems to the world of integers and mathematics, 276 00:12:08,509 --> 00:12:10,389 we need something else. And that something else 277 00:12:10,389 --> 00:12:13,169 appears to be language or at least symbolic representation. 278 00:12:13,629 --> 00:12:15,870 So we've hit the limits of our innate toolkit. 279 00:12:16,230 --> 00:12:19,830 To get to concepts like seven or infinity, we 280 00:12:19,830 --> 00:12:22,610 need language. Now there was this really strong 281 00:12:22,610 --> 00:12:25,850 view associated with Quine that language basically 282 00:12:25,850 --> 00:12:28,049 creates our whole way of seeing the world, our 283 00:12:28,049 --> 00:12:30,870 ontology. Does the research we're looking at 284 00:12:30,870 --> 00:12:33,610 support that really strong form of linguistic 285 00:12:33,610 --> 00:12:36,230 determinism? No, and not in its strongest form. 286 00:12:36,649 --> 00:12:38,929 The idea that language creates our understanding 287 00:12:38,929 --> 00:12:41,850 of basic things like objects existing over time 288 00:12:41,850 --> 00:12:45,470 or cause and effect. That seems too strong. Core 289 00:12:45,470 --> 00:12:47,769 cognition gives us those foundations innately. 290 00:12:47,909 --> 00:12:49,690 Right, we already established that. However, 291 00:12:50,029 --> 00:12:52,350 language is absolutely critical. The sources 292 00:12:52,350 --> 00:12:54,370 argue we need to distinguish between, let's say, 293 00:12:54,450 --> 00:12:56,950 two levels of linguistic influence. There's a 294 00:12:56,950 --> 00:12:59,049 weaker kind, and then there's this much more 295 00:12:59,049 --> 00:13:01,450 powerful, transformative mechanism. Okay, let's 296 00:13:01,450 --> 00:13:03,169 start with the weaker one. How does language 297 00:13:03,169 --> 00:13:05,009 influence a concept that's already kind of there? 298 00:13:05,250 --> 00:13:07,129 This is where language might increase the salience 299 00:13:07,129 --> 00:13:09,929 or help deploy an existing innate distinction. 300 00:13:10,470 --> 00:13:13,240 Think about singular versus plural. one thing 301 00:13:13,240 --> 00:13:15,559 versus many things. OK, baby seems sensitive 302 00:13:15,559 --> 00:13:18,440 to that difference non -verbally. They do. But 303 00:13:18,440 --> 00:13:20,799 some studies, like Barner's work, found something 304 00:13:20,799 --> 00:13:24,279 interesting. Toddlers around two years old only 305 00:13:24,279 --> 00:13:27,539 start reliably using this one versus many distinction 306 00:13:27,539 --> 00:13:30,259 in non -verbal tasks, like if you ask them to 307 00:13:30,259 --> 00:13:34,000 grab the keys. versus the key from a box after 308 00:13:34,000 --> 00:13:35,659 they start actually using plural markers like 309 00:13:35,659 --> 00:13:38,980 S in their own speech. So learning the word keys 310 00:13:38,980 --> 00:13:42,039 helps them consistently apply the concept of 311 00:13:42,039 --> 00:13:44,559 more than one, even when not speaking. It seems 312 00:13:44,559 --> 00:13:46,899 so. The language highlights the distinction. 313 00:13:47,720 --> 00:13:50,299 It makes an existing conceptual contrast more 314 00:13:50,299 --> 00:13:53,240 available, more robust for problem solving. The 315 00:13:53,240 --> 00:13:55,720 concept was there innately, but the linguistic 316 00:13:55,720 --> 00:13:57,659 label sharpens it, brings it to the foreground. 317 00:13:57,879 --> 00:13:59,899 That's like tuning up an existing engine. But 318 00:13:59,899 --> 00:14:02,019 then there's the big one. Quinian bootstrapping. 319 00:14:02,100 --> 00:14:04,080 This is supposed to build entirely new engines, 320 00:14:04,360 --> 00:14:06,000 right? Concepts the old system couldn't even 321 00:14:06,000 --> 00:14:08,620 imagine. That's the idea. This is presented as 322 00:14:08,620 --> 00:14:11,220 a uniquely human ability, the really strong mechanism 323 00:14:11,220 --> 00:14:14,000 of conceptual change. It's about creating a new 324 00:14:14,000 --> 00:14:16,100 representational system. Let's call it Conceptual 325 00:14:16,100 --> 00:14:19,419 System 2, or CS2, that is qualitatively different 326 00:14:19,419 --> 00:14:21,559 from and more powerful than the initial core 327 00:14:21,559 --> 00:14:24,059 cognition system, CS1. And you used the term 328 00:14:24,059 --> 00:14:27,320 incoherent before. The new concepts can be incoherent 329 00:14:27,309 --> 00:14:30,370 from the old perspective? Yes, that's a key aspect. 330 00:14:30,789 --> 00:14:35,029 It means the new system, CS2, might rely on distinctions 331 00:14:35,029 --> 00:14:37,370 or principles that literally cannot be expressed 332 00:14:37,370 --> 00:14:41,009 or represented within the old system, CS1. Consider 333 00:14:41,009 --> 00:14:43,570 a young child's early concept of heaviness. OK. 334 00:14:43,870 --> 00:14:46,029 Initially, they might conflate what adults distinguish 335 00:14:46,029 --> 00:14:48,950 as weight and density. A big fluffy pillow is 336 00:14:48,950 --> 00:14:51,210 heavy, and a small metal weight is heavy. They 337 00:14:51,210 --> 00:14:53,230 don't grasp why one floats and the other sinks. 338 00:14:53,730 --> 00:14:56,750 The adult concept of density. mass per unit volume 339 00:14:56,750 --> 00:14:59,730 requires relating weight and size in a way their 340 00:14:59,730 --> 00:15:02,590 initial undifferentiated heaviness concept simply 341 00:15:02,590 --> 00:15:04,990 doesn't support. From the perspective of that 342 00:15:04,990 --> 00:15:07,409 simple heaviness concept, density is kind of 343 00:15:07,409 --> 00:15:09,990 incoherent. It involves variables the child isn't 344 00:15:09,990 --> 00:15:12,409 tracking yet. So to build that new concept of 345 00:15:12,409 --> 00:15:14,929 density or concepts like exact number, you need 346 00:15:14,929 --> 00:15:17,590 something CS1 doesn't provide. Precisely. And 347 00:15:17,590 --> 00:15:20,090 this is where language or other symbolic systems 348 00:15:20,090 --> 00:15:22,269 are proposed to come in. They provide placeholder 349 00:15:22,269 --> 00:15:25,100 structures. Placeholder structures. Like what? 350 00:15:25,240 --> 00:15:28,360 Think of the ordered count list. One, two, three, 351 00:15:28,360 --> 00:15:31,899 four. Initially, for a young child, these are 352 00:15:31,899 --> 00:15:34,179 just sounds in a specific order. They don't necessarily 353 00:15:34,179 --> 00:15:36,799 know what four means in terms of quantity yet. 354 00:15:36,960 --> 00:15:39,399 Ah, OK. The meaning at first comes purely from 355 00:15:39,399 --> 00:15:41,419 the structure. Two comes after one, three comes 356 00:15:41,419 --> 00:15:44,340 after two. It's this external ordered symbolic 357 00:15:44,340 --> 00:15:47,899 list that acts as a scaffold. The child can learn 358 00:15:47,899 --> 00:15:50,460 and manipulate the structure before fully grasping 359 00:15:50,460 --> 00:15:52,940 the numerical concepts it will eventually represent. 360 00:15:53,149 --> 00:15:55,809 This symbolic scaffolding allows us to build 361 00:15:55,809 --> 00:15:58,230 systems for concepts like division, negative 362 00:15:58,230 --> 00:16:00,990 numbers, or the infinite integers things utterly 363 00:16:00,990 --> 00:16:03,850 outside the scope of core cognition. So the symbols 364 00:16:03,850 --> 00:16:06,950 provide the framework to hang these new initially 365 00:16:06,950 --> 00:16:10,039 abstract ideas on. That's the core idea of Quinian 366 00:16:10,039 --> 00:16:12,360 bootstrapping. The symbols let you represent 367 00:16:12,360 --> 00:16:14,539 relationships and structures that you can't directly 368 00:16:14,539 --> 00:16:17,559 perceive or represent with analogiconic systems 369 00:16:17,559 --> 00:16:20,100 alone. And this interaction between innate abilities 370 00:16:20,100 --> 00:16:22,580 and language isn't just for numbers, right? It 371 00:16:22,580 --> 00:16:24,679 applies to basic categories, too, like knowing 372 00:16:24,679 --> 00:16:27,220 a duck is a kind of thing. It's complex there, 373 00:16:27,419 --> 00:16:30,879 too. The research suggests infants, even by 12 374 00:16:30,879 --> 00:16:34,039 months, do seem to distinguish kinds of things 375 00:16:34,039 --> 00:16:37,000 like duck, a sortle down, from simple properties 376 00:16:37,000 --> 00:16:40,450 like read an adjective. They use kind membership 377 00:16:40,450 --> 00:16:43,929 to track objects. So Quine's idea that language 378 00:16:43,929 --> 00:16:46,570 creates the kind property distinction seems wrong. 379 00:16:46,789 --> 00:16:49,070 But language still helps. It definitely helps 380 00:16:49,070 --> 00:16:51,629 in learning specific kinds. If you label different 381 00:16:51,629 --> 00:16:54,470 looking toy animals all as blickets, poddlers 382 00:16:54,470 --> 00:16:56,629 are quicker to group them based on their shared 383 00:16:56,629 --> 00:16:59,070 kind or shape. If you use an adjective like, 384 00:16:59,070 --> 00:17:01,590 look, the fuzzy one, they're more likely to group 385 00:17:01,590 --> 00:17:04,130 based on that property, fuzziness. So language 386 00:17:04,130 --> 00:17:06,650 doesn't invent the idea of kind, but it powerfully 387 00:17:06,650 --> 00:17:08,789 directs attention and helps children figure out 388 00:17:08,789 --> 00:17:10,670 which kinds of properties are important in their 389 00:17:10,670 --> 00:17:13,089 culture. Let's really focus now on the prime 390 00:17:13,089 --> 00:17:15,470 example of this bootstrapping process, learning 391 00:17:15,470 --> 00:17:18,789 the positive integers 1, 2, 3, and so on. You 392 00:17:18,789 --> 00:17:21,410 said this new system, the integer list, CS2, 393 00:17:21,769 --> 00:17:25,069 is discontinuous with core cognition. CS1. Why 394 00:17:25,069 --> 00:17:27,170 is that leap so hard? What exactly is missing 395 00:17:27,170 --> 00:17:29,970 from CS1? The fundamental discontinuity lies 396 00:17:29,970 --> 00:17:32,230 in representing the successor function. Remind 397 00:17:32,230 --> 00:17:34,309 us what that is. It's the core principle of counting. 398 00:17:34,890 --> 00:17:37,230 For any natural number n, there is always a unique 399 00:17:37,230 --> 00:17:40,170 next number, n plus 1, 5 follows 4, 6 follows 400 00:17:40,170 --> 00:17:43,930 5, and crucially, this process never ends. It 401 00:17:43,930 --> 00:17:46,769 implies infinity. And our core systems can't 402 00:17:46,769 --> 00:17:48,950 handle that. Neither of them. The object file 403 00:17:48,950 --> 00:17:52,130 system is stuck at 3 or 4. The analog magnitude 404 00:17:52,130 --> 00:17:55,089 system represents numbers as fuzzy estimations 405 00:17:55,089 --> 00:17:57,829 on a sort of mental line. It gives you more or 406 00:17:57,829 --> 00:18:01,529 less, but not the precise, discrete step of exactly 407 00:18:01,529 --> 00:18:04,410 one more, repeated indefinitely. There's no way 408 00:18:04,410 --> 00:18:06,950 to build the concept of an infinite ordered sequence 409 00:18:06,950 --> 00:18:08,990 from those starting points alone. Which explains 410 00:18:08,990 --> 00:18:11,230 why learning to count isn't trivial for kids. 411 00:18:11,509 --> 00:18:13,990 If it were just mapping words onto existing concepts, 412 00:18:14,150 --> 00:18:16,769 it should be fast. Exactly. But it takes years, 413 00:18:16,890 --> 00:18:19,349 typically from age two to four or even five. 414 00:18:19,730 --> 00:18:21,809 Kids spend a long time just mastering the sequence 415 00:18:21,809 --> 00:18:23,490 and then figuring out what the words actually 416 00:18:23,490 --> 00:18:26,829 mean. This prolonged difficulty is strong evidence 417 00:18:26,829 --> 00:18:29,509 that they aren't just labeling old ideas. They're 418 00:18:29,509 --> 00:18:31,710 constructing something fundamentally new. So 419 00:18:31,710 --> 00:18:33,819 how does this construction happen? What are the 420 00:18:33,819 --> 00:18:35,799 steps in this bootstrapping process for number? 421 00:18:36,140 --> 00:18:38,900 It seems to follow a predictable pattern, often 422 00:18:38,900 --> 00:18:41,799 called the subset knower stage. First, as we 423 00:18:41,799 --> 00:18:43,799 mentioned, children learn the count list, one, 424 00:18:43,819 --> 00:18:46,559 two, three, four, five, mostly as a meaningless 425 00:18:46,559 --> 00:18:49,359 chant and ordered sequence of sounds. Like reciting 426 00:18:49,359 --> 00:18:51,900 the alphabet before knowing the letters? Precisely. 427 00:18:52,180 --> 00:18:55,259 That list... is the crucial external placeholder 428 00:18:55,259 --> 00:18:57,599 structure provided by their culture and language. 429 00:18:58,299 --> 00:19:00,299 Then they start to connect the first few words 430 00:19:00,299 --> 00:19:03,039 in that list to actual quantities. They become 431 00:19:03,039 --> 00:19:06,160 subset knowers. A child might first become a 432 00:19:06,160 --> 00:19:08,640 one knower. If you ask them, give me one fish, 433 00:19:08,839 --> 00:19:11,299 they could do it reliably. But ask for two and 434 00:19:11,299 --> 00:19:14,160 they just grab a random handful. Then maybe months 435 00:19:14,160 --> 00:19:16,539 later, they become a two knower. They can give 436 00:19:16,539 --> 00:19:19,259 you one and they can give you exactly two. But 437 00:19:19,259 --> 00:19:21,799 ask for three and they're back to grabbing randomly. 438 00:19:21,930 --> 00:19:24,269 They learn the meanings of these first few number 439 00:19:24,269 --> 00:19:27,269 words one by one by one. And how are they grounding 440 00:19:27,269 --> 00:19:30,670 the meaning of, say, one or two? It seems they 441 00:19:30,670 --> 00:19:33,650 map these first few number words onto the representations 442 00:19:33,650 --> 00:19:36,589 provided by their parallel individuation object 443 00:19:36,589 --> 00:19:40,309 file system. One maps to a mental file for a 444 00:19:40,309 --> 00:19:43,630 single individual eye. Two maps to a model with 445 00:19:43,630 --> 00:19:48,289 two individuals, I1, I2. Three maps to I1, I2, 446 00:19:48,289 --> 00:19:52,529 I3. They use the exact tracking system. CS1, 447 00:19:52,910 --> 00:19:54,730 to give meaning to the first few symbols from 448 00:19:54,730 --> 00:19:57,289 the list, CS2. But they hit the wall around three 449 00:19:57,289 --> 00:19:59,410 or four, just like the object file system limit. 450 00:19:59,650 --> 00:20:01,930 Exactly. They might become a three knower, maybe 451 00:20:01,930 --> 00:20:04,009 even a four knower, but they're still stuck. 452 00:20:04,349 --> 00:20:06,670 They haven't grasped the underlying system of 453 00:20:06,670 --> 00:20:09,029 the count list yet. The real breakthrough is 454 00:20:09,029 --> 00:20:11,170 what's called the inductive leap, or acquiring 455 00:20:11,170 --> 00:20:13,089 the cardinal principle. The cardinal principle. 456 00:20:13,329 --> 00:20:15,130 That sounds important. What clicks at that moment? 457 00:20:15,349 --> 00:20:18,170 It's the aha moment. The child suddenly generalizes. 458 00:20:18,509 --> 00:20:20,250 They realize that the last number of words said 459 00:20:20,250 --> 00:20:22,329 when counting a set represents the total number 460 00:20:22,329 --> 00:20:24,990 of items in that set, the cardinality. And crucially, 461 00:20:25,289 --> 00:20:27,250 they seem to grasp implicitly that every word 462 00:20:27,250 --> 00:20:29,789 on the count list refers to a quantity exactly 463 00:20:29,789 --> 00:20:31,730 one greater than the previous word they get the 464 00:20:31,730 --> 00:20:34,190 successor function. So they realize the whole 465 00:20:34,190 --> 00:20:36,730 list has meaning, not just the first few words. 466 00:20:36,970 --> 00:20:39,750 Yes. The placeholder structure isn't just sounds 467 00:20:39,750 --> 00:20:42,390 anymore. It's suddenly understood as a generative 468 00:20:42,390 --> 00:20:44,940 rule for representing number. They might not 469 00:20:44,940 --> 00:20:47,359 be able to count perfectly to 100 yet, but they 470 00:20:47,359 --> 00:20:49,500 understand the principle that allows counting 471 00:20:49,500 --> 00:20:52,299 to continue indefinitely. They bootstrapped the 472 00:20:52,299 --> 00:20:55,420 integer concept, CS2. Okay, so now they have 473 00:20:55,420 --> 00:20:58,500 this abstract, powerful number system in their 474 00:20:58,500 --> 00:21:01,259 heads, but does that mean they instantly feel 475 00:21:01,259 --> 00:21:04,559 the difference between, say, 7 and 10? Does it 476 00:21:04,559 --> 00:21:07,549 connect back to that gut sense of quantity? Not 477 00:21:07,549 --> 00:21:09,549 necessarily. And this is another fascinating 478 00:21:09,549 --> 00:21:11,910 piece of evidence for discontinuity. The next 479 00:21:11,910 --> 00:21:15,049 step is mapping this new abstract integer list, 480 00:21:15,250 --> 00:21:18,690 CS2, back onto the old approximate analog magnitude 481 00:21:18,690 --> 00:21:21,210 system, CS1. Why do they need to do that? To 482 00:21:21,210 --> 00:21:23,130 get quantitative intuition. To know roughly how 483 00:21:23,130 --> 00:21:25,809 much 7 is compared to 10. And research shows 484 00:21:25,809 --> 00:21:27,369 that children who have just made the cardinal 485 00:21:27,369 --> 00:21:30,269 principle leap, so -called CP knowers, are often 486 00:21:30,269 --> 00:21:32,450 initially non -mappers. Non -mappers. They know 487 00:21:32,450 --> 00:21:34,539 the counting rule. They can tell you that seven 488 00:21:34,539 --> 00:21:36,519 comes after six. They might be able to count 489 00:21:36,519 --> 00:21:38,900 out seven items correctly, but if you show them 490 00:21:38,900 --> 00:21:41,759 two sets of dots, say seven versus ten, and ask 491 00:21:41,759 --> 00:21:46,230 which is more, they might guess randomly or rely 492 00:21:46,230 --> 00:21:48,509 on visual cues that are subject to Weber's law, 493 00:21:48,869 --> 00:21:50,849 just like a pre -counting child would. So they 494 00:21:50,849 --> 00:21:53,329 have the abstract rule, but not the connection 495 00:21:53,329 --> 00:21:56,009 to magnitude for those larger numbers yet. Exactly. 496 00:21:56,349 --> 00:21:58,769 They haven't yet associated the specific symbol 497 00:21:58,769 --> 00:22:01,390 seven with the corresponding fuzzy magnitude 498 00:22:01,390 --> 00:22:04,490 on their internal number line. That mapping process, 499 00:22:04,829 --> 00:22:07,029 connecting the precise symbols back to the approximate 500 00:22:07,029 --> 00:22:10,769 sense of quantity, is slow and happens gradually 501 00:22:10,769 --> 00:22:13,069 after the cardinal principle is acquired. And 502 00:22:13,069 --> 00:22:15,750 that slowness is key evidence. It's powerful 503 00:22:15,750 --> 00:22:18,269 evidence. If seven were just a continuous extension 504 00:22:18,269 --> 00:22:20,829 of the core concept of three, the mapping should 505 00:22:20,829 --> 00:22:23,609 be automatic. The fact that it's laborious and 506 00:22:23,609 --> 00:22:25,789 piecemeal strongly suggests that the integer 507 00:22:25,789 --> 00:22:29,200 concept was constructed discontinuously. using 508 00:22:29,200 --> 00:22:31,380 the linguistic list as a necessary scaffold, 509 00:22:31,799 --> 00:22:34,279 creating a system that then needs effortful integration 510 00:22:34,279 --> 00:22:36,559 with the older core systems. This really brings 511 00:22:36,559 --> 00:22:39,059 home the power of symbolic systems. And it makes 512 00:22:39,059 --> 00:22:40,880 you think about cultures that might lack these 513 00:22:40,880 --> 00:22:43,039 extensive number systems, like the Pirahan people 514 00:22:43,039 --> 00:22:45,460 in the Amazon. The Pirahan case is often cited 515 00:22:45,460 --> 00:22:48,140 here, yes. Reports suggest their language lacks 516 00:22:48,140 --> 00:22:51,099 number words beyond maybe one, two, and many. 517 00:22:51,250 --> 00:22:54,410 And studies testing their ability to mask quantities 518 00:22:54,410 --> 00:22:57,490 exactly for sets larger than three show their 519 00:22:57,490 --> 00:22:59,890 performance immediately drops off and looks just 520 00:22:59,890 --> 00:23:02,210 like the signature of the analog magnitude system, 521 00:23:02,529 --> 00:23:04,829 approximate and ratio dependent. So without the 522 00:23:04,829 --> 00:23:07,170 symbolic tool, the count list, they seem to rely 523 00:23:07,170 --> 00:23:09,869 solely on the core cognition systems for quantity. 524 00:23:10,109 --> 00:23:12,589 That's the interpretation. It provides a stark 525 00:23:12,589 --> 00:23:15,500 illustration. The development of exact large 526 00:23:15,500 --> 00:23:17,779 number concepts seems deeply reliant on having 527 00:23:17,779 --> 00:23:20,579 that specific linguistic or symbolic scaffolding 528 00:23:20,579 --> 00:23:23,019 in place. That's the interpretation. It provides 529 00:23:23,019 --> 00:23:26,079 a stark illustration. Hashtag tag outro. So pulling 530 00:23:26,079 --> 00:23:28,660 all this together, what does it tell us? It paints 531 00:23:28,660 --> 00:23:31,039 a picture of the human mind, starting with a 532 00:23:31,039 --> 00:23:34,380 surprisingly rich foundation. These innate core 533 00:23:34,380 --> 00:23:37,059 cognition systems for understanding objects, 534 00:23:37,420 --> 00:23:40,400 agents, space, and approximate number. They give 535 00:23:40,400 --> 00:23:43,839 us continuous intuitive grasp of the basic structure 536 00:23:43,839 --> 00:23:45,640 of the world. But that's not the whole story. 537 00:23:45,880 --> 00:23:48,900 Not at all. To get to the truly powerful abstract 538 00:23:48,900 --> 00:23:51,900 concepts that define human thought, formal mathematics, 539 00:23:52,619 --> 00:23:55,799 science, complex planning, we need something 540 00:23:55,799 --> 00:23:58,670 more. We need this remarkable process. quitting 541 00:23:58,670 --> 00:24:01,089 and bootstrapping. It uses language or other 542 00:24:01,089 --> 00:24:04,170 symbolic systems as essential placeholders. Scaffolding 543 00:24:04,170 --> 00:24:06,250 to build something new. Exactly. Scaffolding 544 00:24:06,250 --> 00:24:07,849 that allows the mind to make these inductive 545 00:24:07,849 --> 00:24:10,650 leaps, constructing new conceptual systems, CS2, 546 00:24:10,710 --> 00:24:12,609 that are qualitatively different, more powerful, 547 00:24:12,890 --> 00:24:14,829 and sometimes from the perspective of the original 548 00:24:14,829 --> 00:24:18,349 system, CS1, even seem incoherent initially. 549 00:24:18,569 --> 00:24:21,289 Yeah, that idea of incoherence is really striking. 550 00:24:21,950 --> 00:24:24,289 Learning isn't just adding information. It can 551 00:24:24,289 --> 00:24:26,849 be about fundamentally restructuring your way 552 00:24:26,849 --> 00:24:29,769 of thinking, like the child moving from simple 553 00:24:29,769 --> 00:24:32,950 heaviness to distinguishing weight and density, 554 00:24:33,309 --> 00:24:36,190 or from tracking three items to understanding 555 00:24:36,190 --> 00:24:39,220 the infinite integers. It's hard work, because 556 00:24:39,220 --> 00:24:41,339 the new framework might not even make sense from 557 00:24:41,339 --> 00:24:44,119 the old point of view. It is hard work. And it 558 00:24:44,119 --> 00:24:46,240 highlights how the very structure of the symbolic 559 00:24:46,240 --> 00:24:48,240 systems we learn the grammar of our language, 560 00:24:48,339 --> 00:24:51,099 the notation of mathematics, shapes the conceptual 561 00:24:51,099 --> 00:24:54,079 landscape we can build. These tools don't just 562 00:24:54,079 --> 00:24:56,480 express thoughts. They seem to enable certain 563 00:24:56,480 --> 00:24:58,480 kinds of thoughts in the first place. Which leaves 564 00:24:58,480 --> 00:25:00,579 us with a really provocative final thought, doesn't 565 00:25:00,579 --> 00:25:03,119 it? If the specific structure of language, like 566 00:25:03,119 --> 00:25:05,940 having a count list, is necessary to bootstrap 567 00:25:05,940 --> 00:25:08,890 a concept as basic as exact number. Well, it 568 00:25:08,890 --> 00:25:11,130 makes you wonder, right? Think about the language 569 00:25:11,130 --> 00:25:13,490 or languages you use every day. What complex 570 00:25:13,490 --> 00:25:16,609 ideas may be about time or consciousness or causality 571 00:25:16,609 --> 00:25:19,430 or perhaps dimensions beyond our immediate perception? 572 00:25:20,150 --> 00:25:22,230 What concepts might currently be hovering just 573 00:25:22,230 --> 00:25:24,509 beyond our conceptual reach? Because we lack 574 00:25:24,509 --> 00:25:27,650 the right words, the right symbols. Perhaps. 575 00:25:28,140 --> 00:25:31,180 Because we haven't yet collectively built the 576 00:25:31,180 --> 00:25:33,359 necessary symbolic scaffolding, the placeholder 577 00:25:33,359 --> 00:25:35,140 structures that would allow our minds to make 578 00:25:35,140 --> 00:25:38,140 the next great inductive leap, what potential 579 00:25:38,140 --> 00:25:40,599 conceptual systems remain unbuilt, waiting for 580 00:25:40,599 --> 00:25:41,640 the right linguistic tools?