Darwin’s Top 10 Arguments Against His Own Theory

Herman B. Bouma https://evolutionnews.org/2023/03/darwins-top-10-arguments-against-his-own-theory/ TYhe purpose of this post is not to propose contemporary arguemnts to criticize evolution. It is to illustrate how a balenced, objective examination of the theory might look. Both sides of the vociferous contemporarty debate could learn from this example.] In a recent article here I referred to my canceled presentation at the annual conference of the National Science Teaching Association. My topic was to be, “The Top 10 Scientific Arguments Against Darwin’s Theory — According to Darwin Himself.” What are those top ten arguments? Let’s take a look. Charles Darwin took seriously objections to his theory that had been raised by many of the most eminent naturalists of his day. In The Origin of Species he considered in detail 37 of them. Darwin acknowledged that there were “a crowd of difficulties” with his theory and stated, “Some of them are so serious that to this day I can hardly reflect on them without being in some degree staggered; but, to the best of my judgment, the greater number are only apparent, and those that are real are not, I think, fatal to the theory.” (p. 158) (All citations are to Charles Darwin, The Origin of Species (New York: NAL Penguin Inc., 1958).) Based on Darwin’s discussion in The Origin of Species, it is reasonable to conclude that he considered the arguments set forth below to be the top ten scientific arguments against his theory. These arguments relate to Darwin’s proposed mechanism for evolution, i.e., the application of natural selection to randomly produced variations. (Note: With respect to the origin of life, Darwin theorized in The Origin of Species that the very first forms of life (at most, eight to ten forms) were produced by the Creator. He did not consider any arguments against this part of his theory and apparently none were raised by the naturalists of his day, most of whom subscribed to a theory of design.) Darwin’s thoughtful consideration of the scientific arguments against his theory is consistent with his hope for the future: “I look with confidence to the future, — to young and rising naturalists, who will be able to view both sides of the question with impartiality.” (p. 444) 1. The Complexity of Eyes Darwin states, “To suppose that the eye with all its inimitable contrivances for adjusting the focus to different distances, for admitting different amounts of light, and for the correction of spherical and chromatic aberration, could have been formed by natural selection, seems, I freely confess, absurd in the highest degree.” (p. 168) 2. Existence of Similar Organs in Remotely Allied Species Darwin considers the electric organs of species of fish that are remotely allied and also the luminous organs of insects “belonging to widely different families.” He states that the latter “offer, under our present state of ignorance, a difficulty almost exactly parallel with that of the electric organs.” (p. 176) Darwin admits the difficulty under his theory “of an organ, apparently the same, arising in several remotely allied species.” (p. 176) 3. Existence of Different Organs for the Same Function in Closely Allied Species Darwin considers two genera of orchid, the Coryanthes and the Catasetum. He explains in detail the ingenious “contrivance” that the Coryanthes uses for pollination. He then turns to the Catasetum, which is “closely allied” to the Coryanthes, and states that the construction of the flower in the Catasetum “is widely different, though serving the same end.” (p. 180) Darwin admits that it is common throughout nature for the same end to be gained by the most diversified means, “even sometimes in the case of closely-related beings.” (p. 178) 4. Parts with Little Importance Darwin states, “I have sometimes felt great difficulty in understanding the origin or formation of parts of little importance; almost as great, though of a very different kind, as in the case of the most perfect and complex organs.” (p. 181) He mentions the tail of the giraffe as an example of a part with little apparent importance. He states that it looks like “an artificially constructed fly-flapper” and “[it] seems at first incredible that this could have been adapted for its present purpose by successive slight modifications, each better and better fitted, for so trifling an object as to drive away flies.” (p. 181) 5. Complex Instincts Darwin acknowledges, “Many instincts are so wonderful that their development will probably appear to the reader a difficulty sufficient to overthrow my whole theory.” (p. 228) Darwin states, “He must be a dull man who can examine the exquisite structure of a [honeycomb], so beautifully adapted to its end, without enthusiastic admiration. We hear from mathematicians that bees have practically solved a recondite problem, and have made their cells of the proper shape to hold the greatest possible amount of honey, with the least possible consumption of precious wax in their construction. It has been remarked that a skilful workman … would find it very difficult to make cells of wax of the true form, though this is effected by a crowd of bees working in a dark hive.” (pp. 242-243) 6. Neuter Ants and Their Different Castes With respect to neuter ants, Darwin states it is “one special difficulty, which at first appeared to me insuperable, and actually fatal to the whole theory … for these neuters often differ widely in instinct and in structure from both the males and fertile females, … yet, from being sterile, they cannot propagate their kind.” (p. 250) Darwin goes on to state, “But we have not as yet touched on the acme of the difficulty; namely, the fact that the neuters of several ants differ, not only from the fertile females and males, but from each other, sometimes to an almost incredible degree, and are thus divided into two or even three castes.” (p. 253) He acknowledges, “It will indeed be thought that I have an overweening confidence in the principle of natural selection, when I do not admit that such wonderful and well-established facts at once annihilate the theory.” (p. 253) 7. The Eyes of the Flat-Fish During its early youth the body of the flat-fish is symmetrical with one eye on each side. However, as the body matures, one eye “begins to glide slowly round the head” to the other side. (pp. 209-210) This is beneficial because the adult flat-fish spends most of its time lying on its side on the bottom of the ocean. Darwin agrees that his theory of natural selection cannot account for this feature. He states that it “may be attributed to the habit, no doubt beneficial to the individual and to the species, of endeavouring to look upwards with both eyes, whilst resting on one side at the bottom.” (p. 211) Thus, it “may be attributed almost wholly to continued use, together with inheritance.” (pp. 222-223) 8. Absence of Transitional Forms in the Fossil Record With respect to the absence of transitional forms in the fossil record, Darwin states that under his theory, “…as this process of extermination has acted on an enormous scale, so must the number of intermediate varieties, which have formerly existed, be truly enormous. Why then is not every geological formation and every stratum full of such intermediate links? Geology assuredly does not reveal any such finely-graduated organic chain; and this, perhaps, is the most obvious and serious objection which can be urged against the theory.” (p. 287) Darwin admits that “though we do find many links — we do not find interminable varieties, connecting together all extinct and existing forms by the finest graduated steps.” (pp. 335-336) He states, “That the geological record is imperfect all will admit; but that it is imperfect to the degree required by our theory, few will be inclined to admit.” (p. 431) He also acknowledges, “He who rejects this view of the imperfection of the geological record, will rightly reject the whole theory.” (p. 336) 9. Absence of Transitional Forms Even Within Particular Geological Formations With respect to the absence of transitional forms even within particular geological formations, Darwin states, “[I]t cannot be doubted that the geological record, viewed as a whole, is extremely imperfect; but if we confine our attention to any one formation, it becomes much more difficult to understand why we do not therein find closely graduated varieties between the allied species which lived at its commencement and at its close.” (p. 298) He confesses, “But I do not pretend that I should ever have suspected how poor was the record in the best preserved geological sections, had not the absence of innumerable transitional links between the species which lived at the commencement and close of each formation, pressed so hardly on my theory.” (pp. 304-305) 10. Sudden Appearance of New Forms of Life Darwin states, “The abrupt manner in which whole groups of species suddenly appear in certain formations, has been urged by several paleontologists — for instance, by Agassiz, Pictet, and Sedgwick — as a fatal objection to the belief in the transmutation of species.” (p. 305) He goes on to state, “There is another and allied difficulty, which is much more serious. I allude to the manner in which species belonging to several of the main divisions of the animal kingdom suddenly appear in the lowest known fossiliferous rocks,” i.e., the Cambrian strata. (p. 308) “To the question why we do not find rich fossiliferous deposits … prior to the Cambrian system, I can give no satisfactory answer.” (p. 309) Darwin concludes, “The case at present must remain inexplicable; and may be truly urged as a valid argument against the views here entertained.” (p. 310)

Examples of chatgpt "reasoning" and "intelligence"

If the "intelligence" of chatgpt has you worried for the future of the world, consider theese real interactions with chatgpt: The four-color theorem states that any map in a plane can be colored using four-colors in such a way that regions sharing a common boundary (other than a single point) do not share the same color. Is there are map - not in a plane - that requires five colors? Yes, there are maps that require five or more colors. The four-color theorem only applies to maps in a plane. Give an example. An example of a map that requires five colors is a map of the United States where states are considered as regions and neighboring states must be colored differently. This map cannot be colored using only four colors as there are states, such as Montana and North Dakota, that share a common boundary but are adjacent to different states and therefore cannot share the same color. But a map of the United States is a map in a plane and the four color theorem states that such a map can be colored with four colors. My apologies, you are correct. A map of the United States can be colored using only four colors. The example I provided does not contradict the four-color theorem. Thank you for pointing this out. So do you still maintain that there are maps - not in a plane - that require five colors? Yes, maps that are not in a plane, such as maps of a sphere or a torus, can require five or more colors. In such cases, the four-color theorem does not apply. Good. Can you give an exact example? Sure! An example of a map that requires five colors is a map of a cube. Each face of a cube can be considered as a region and each vertex of the cube is a point where three faces meet. In order to properly color such a map, each face must be colored differently from its adjacent faces, which requires at least five colors. The four color theorem specifies regions sharing a common boundary (other than a single point) . A vertex is a single point. So a vertex is ruled out. I apologize for the confusion. You are correct, a vertex is a single point and the four-color theorem does not apply to single points. In the case of a cube map, each face can be considered as a region and they share edges, which are common boundaries, and thus the four-color theorem applies. In this case, four colors would be sufficient to properly color the map of the cube. Thank you for pointing out the error in my previous answer. If someone goes due south for 50 miles, then due west for 50 miles, and finally due north for 50 miles, can he end up at his starting point? No, the person cannot end up at their starting point. But there are examples that it is possible to end up at the starting point. You are correct. If a person goes due south for 50 miles, then due west for 50 miles, and finally due north for 50 miles, it is possible for them to end up at their starting point if they start and end at the North Pole. This is because, at the North Pole, all directions are south, and any displacement to the south will result in the person moving closer to the equator. However, in most other locations, this trajectory would not result in the person returning to their starting point. there are other examples. Yes, there are other examples where a person going due south for 50 miles, then due west for 50 miles, and finally due north for 50 miles can end up at their starting point. These would occur in locations where the east-west distance covered in the westward leg of the journey is exactly equal to the north-south displacement in the southward and northward legs. In such cases, the net effect of the journey would be zero, and the person would end up back at their starting point. Rubbish I'm sorry if my response was not helpful or accurate. Please let me know what specific information you are looking for and I will do my best to assist you. I gave you the answer in a different chat I apologize for the confusion. As a language model AI, I don't have access to previous conversations or context across multiple interactions. Could you please restate your question or provide the information you would like to know more about? I would be happy to help. All cows are animal. Are all non-animals non-cows? No, not all non-animals are non-cows. Some objects or substances that are not animals, such as plants or rocks, could be considered non-cows, but others like buildings or machines, are not considered either animals or cows. Your answer is nonsensical. To show that not all non-animals are non-cows, you need to find something that is a non-animal but is not a non-cow. You have not done that. I apologize for the confusion. You are correct. All non-animals are non-cows by definition, since cows are a specific type of animal. Thank you for bringing this to my attention. Do you know what a contrapositive is? Yes, I'm familiar with contrapositive. In logic, the contrapositive of a proposition is the negation of its converse. The contrapositive of a statement of the form "If A then B" is "If not B then not A." The contrapositive is logically equivalent to the original statement, meaning that if the original statement is true, then its contrapositive is also true, and vice versa. The contrapositive is often used to prove the validity of a statement or argument by showing that its contrapositive is logically sound. Is what you just wrote related to my original question? Yes, it is related to your original question. The contrapositive of the statement "All cows are animals" is "If something is not an animal, then it is not a cow." This contrapositive statement is logically equivalent to the original statement, meaning that if the original statement is true, then its contrapositive is also true, and vice versa. In this case, the contrapositive is useful for confirming that all non-animals are indeed non-cows. -- Is "this sentence is true" true? This statement is a paradox and is known as the Liar Paradox. It creates a self-contradiction, as the sentence itself cannot be both true and false at the same time. Read the sentence again - you substituted "false" for "true" - you followed your expectations and did not read carefully! My apologies, let me correct my mistake. The statement "This sentence is true" creates a paradox and is known as the Liar Paradox. It creates a self-contradiction, as the sentence itself cannot be both true and false at the same time. Not at all! That would be correct if I had written "this sentence is false". But I wrote "this sentence is true". Try again. If you are impressed with or worried about the "intelligence" of chatgpt taking over the world, considdr these examples of its logical powers. The statement "This sentence is true" is true. Why? What makes it true" The statement "This sentence is true" is true by definition. By claiming to be true, it satisfies the definition of being true, making it a true statement. This creates a self-referential loop where the truth of the statement is established by its own assertion of truth. I am a professor of philosophy. You do not get a passing grade. I apologize if my previous answer did not fully meet your expectations. As an AI language model, I aim to provide accurate and relevant information, but I am not perfect and can make mistakes. Please let me know how I can better assist you.

New suggested readings

I have been unable to open my bnlog for monmths. Suddenly I am in! So here are some new suggested radimngs. https://evolutionnews.org/2023/03/peer-reviewed-paper-neo-darwinism-must-mutate-to-survive/ https://evolutionnews.org/2022/10/the-human-fossil-record-lacks-intermediaries/ https://evolutionnews.org/2022/10/do-fossils-demonstrate-human-evolution-lets-consider-at-the-technical-literature/

Archeology vs the Torah

 

Eruvin, Archeology and the Conquest of Canaan

Posted by: Gil Student in From The Hashkafah Files, Magazine, Musings, Posts Jul 8, 22 0

by R. Gil Student

I. The Conquest of Canaan

Many historians with a critical eye toward the Bible point out that the archeological record seems to contradict the biblical account of Yehoshua’s conquest of the land. While the archeological record shows that many areas of the land contain artifacts showing evidence of a Jewish presence at that, many places lack that evidence — implying that the Jews did not, in fact, conquer the entire land in the time of Yehoshua. For someone unfamiliar with the Bible and traditional commentaries, this might seem like a significant historical challenge to the accuracy of the Bible. I would like explore a response from a different literature.

R. Amnon Bazak (cont., Israel; To This Very Day: Fundamental Questions in Bible Study, pp. 269, 271), before responding to it, summarizes the problem as follows:

The conquest of the Land of Israel is described at length in the book of Joshua, and the conventional view, based on a superficial reading of the text, is that both the process of conquest and the subsequent settlement of the tribes of Israel in the land were completed in a short time. Various archeological findings cast grave doubts on this view, however… From the evidence we may conclude that the Canaanite centers were not attacked all at the same time, but rather over a long period…

I disagree with R. Bazak on one point here. I don’t think even a superficial reading allows for the conclusion that the entire land was conquered in the seven years of Yehoshua’s wars. The Bible clearly says, “And Yehoshua was old and advanced in age, and God said to him, ‘You are old and advanced in age, yet there remains much of the land to still be possessed’” (Josh. 13:1). The Torah even says about the nations in the land, “Little by little I will drive them out before you, until you grow numerous and inherit the land” (Ex. 23:30). However, since I can be accused of apologetic reading of the text to answer this and similar challenges, I will set that aside and instead respond from a great halakhic debate of the early nineteenth century, in which both participants presumably were unaware of biblical criticism and certainly were unaware of the late twentieth century historical challenges to the Bible.

II. The Great Eruv Debate

Around the year 1814, Rav Aharon of Karlin questioned the validity of the modern eruv. He was a leading Lithuanian rabbi, the grandson of Rav Baruch of Shklov, one of the leading disciples of the Vilna Gaon, and himself among the leading students of Rav Chaim Volozhiner. Rav Aharon of Karlin sent a long letter detailing his questions and criticisms to Rav Ephraim Zalman Margoliyos of Brody, the leading Galician scholar of his day.

An eruv in this context means the rabbinic device established to permit carrying on Shabbos in areas in which carrying would otherwise be rabbinically forbidden. Because this is only a rabbinic innovation, it does not work where carrying is biblically forbidden. Rav Aharon of Karlin writes at length exploring the various reasons to say that carrying in most cities today is forbidden only on a rabbinic level and with great expertise and comprehensive command of the commentarial literature challenges them all. If he is correct, today’s eruvin do not work and carrying is forbidden on Shabbos.

Rav Ephraim of Brody replied at length and with equally impressive expertise, defending the contemporary eruv and questioning Rav Aharon of Karlin’s conclusions. The two continued exchanging letters in what became a celebrated debate that has yet to be resolved. Common practice follows Rav Ephraim but many scholars personally follow Rav Aharon. In a recent book containing the exchange of letters and additional material and footnotes, the editor — Rav Simcha Rotsztein — lists 75 areas of disagreement between the two scholars (Kol Kisvei Ha-Mishkenos Ya’akov Ve-Ha-Beis Ephraim Be-Inyan Asu Rabim U-Mevatlei Mechitzos, Lakewood, 2020, pp. 221-223).

Rav Aharon of Karlin primarily asked questions based on classical rabbinic texts and commentaries. However, he also asked a question based on the biblical narrative. According to the prevailing halakhic theory that Rav Aharon questioned, there can only be a public domain if 600,000 people pass through the area. Exactly how and how often they must pass through is a matter of dispute. However, Rav Aharon challenges this theory entirely, arguing that the vast majority of commentators reject it. He also challenges the theory based on the story of the Jews conquering the land of Israel after wandering in the desert for forty years. If 600,000 Jews entered Israel and then settled in different parts of the land, how could there have been any area in which 600,000 people passed through? Some lived in northern Israel, some in the south, etc. (Mishkenos Ya’akov, no. 108 s.v. ve-gam ki nireh).

III. Eruv and the Slow Conquest of Canaan

Rav Ephraim Zalman Margoliyos of Brody (Beis Ephraim, no. 26 s.v. unsh”k mi-divrei ha-Tosafos) replies that the 600,000 only included the males above the age of 20. There were also women and children. Additionally, there were Erev Rav, gentiles who joined with the Jewish people, as well as converts and servants. There were also merchants from other nations who traveled in the land. And finally, adds Rav Ephraim, the Jews did not conquer and settle the whole land (ve-gam ki lo malu es ha-aretz ve-kavshuha). They defeated 31 kings but there was a lot of land left unconquered (ve-nisharah ba-aretz harbeh me’od). Those gentiles contributed to the number of people passing through the land. Therefore, concludes Rav Ephraim, there were many more than 600,000 people in the land of Israel at that time and there could have been an area where carrying was biblically prohibited according to those who hold that there needs to be 600,000 people passing through. To my knowledge, Rav Aharon of Karlin never returned to this specific point.

We see here, in the midst of a tense debate between two great early nineteenth century halakhic scholars, completely outside the realm of biblical apologetics, the assertion that the conquest of the land of Israel was slow. This is contrary to the challenge of archeologists and historians, who claim that the conquest was fast and the archeological record contradicts the biblical account. Rav Ephraim Zalman Margoliyos of Brody read the biblical account as one of a slow conquest, which supports his halakhic readings of rabbinic texts and medieval Talmud commentaries. Setting aside the question of the sustainability of such a large population at the time, which while important is not our current discussion, the pace of conquest within the halakhic literature matches that of the archeological and historical literature. I do not believe that the halakhic literature needs the support of archeology. However, in this case, the explicit statements in a halakhic context respond to the challenges of archeologists questioning the biblical text.

What Does It Mean for AI to Understand?

It’s simple enough for AI to seem to comprehend data, but devising a true test of a machine’s knowledge has proved difficult.

 https://www.quantamagazine.org/what-does-it-mean-for-ai-to-understand-20211216/?utm_source=Quanta+Magazine&utm_campaign=fffe95a7f6-RSS_Daily_Computer_Science&utm_medium=email&utm_term=0_f0cb61321c-fffe95a7f6-389846569&mc_cid=fffe95a7f6&mc_eid=61275b7d81

Melanie Mitchell

Contributing Columnist

---

December 16, 2021

---

[[More grist for AI skepticism.]]

Remember IBM’s Watson, the AI Jeopardy! champion? A 2010 promotion proclaimed, “Watson understands natural language with all its ambiguity and complexity.” However, as we saw when Watson subsequently failed spectacularly in its quest to “revolutionize medicine with artificial intelligence,” a veneer of linguistic facility is not the same as actually comprehending human language.

Natural language understanding has long been a major goal of AI research. At first, researchers tried to manually program everything a machine would need to make sense of news stories, fiction or anything else humans might write. This approach, as Watson showed, was futile — it’s impossible to write down all the unwritten facts, rules and assumptions required for understanding text. More recently, a new paradigm has been established: Instead of building in explicit knowledge, we let machines learn to understand language on their own, simply by ingesting vast amounts of written text and learning to predict words. The result is what researchers call a language model. When based on large neural networks, like OpenAI’s GPT-3, such models can generate uncannily humanlike prose (and poetry!) and seemingly perform sophisticated linguistic reasoning.

But has GPT-3 — trained on text from thousands of websites, books and encyclopedias — transcended Watson’s veneer? Does it really understand the language it generates and ostensibly reasons about? This is a topic of stark disagreement in the AI research community. Such discussions used to be the purview of philosophers, but in the past decade AI has burst out of its academic bubble into the real world, and its lack of understanding of that world can have real and sometimes devastating consequences. In one study, IBM’s Watson was found to propose “multiple examples of unsafe and incorrect treatment recommendations.” Another study showed that Google’s machine translation system made significant errors when used to translate medical instructions for non-English-speaking patients.

How can we determine in practice whether a machine can understand? In 1950, the computing pioneer Alan Turing tried to answer this question with his famous “imitation game,” now called the Turing test. A machine and a human, both hidden from view, would compete to convince a human judge of their humanness using only conversation. If the judge couldn’t tell which one was the human, then, Turing asserted, we should consider the machine to be thinking — and, in effect, understanding.

Unfortunately, Turing underestimated the propensity of humans to be fooled by machines. Even simple chatbots, such as Joseph Weizenbaum’s 1960s ersatz psychotherapist Eliza, have fooled people into believing they were conversing with an understanding being, even when they knew that their conversation partner was a machine.

In a 2012 paper, the computer scientists Hector Levesque, Ernest Davis and Leora Morgenstern proposed a more objective test, which they called the Winograd schema challenge. This test has since been adopted in the AI language community as one way, and perhaps the best way, to assess machine understanding — though as we’ll see, it is not perfect. A Winograd schema, named for the language researcher Terry Winograd, consists of a pair of sentences, differing by exactly one word, each followed by a question. Here are two examples:

Sentence 1: I poured water from the bottle into the cup until it was full.

Question: What was full, the bottle or the cup?

Sentence 2: I poured water from the bottle into the cup until it was empty.

Question: What was empty, the bottle or the cup?

Sentence 1: Joe’s uncle can still beat him at tennis, even though he is 30 years older.

Question: Who is older, Joe or Joe’s uncle?

Sentence 2: Joe’s uncle can still beat him at tennis, even though he is 30 years younger.

Question: Who is younger, Joe or Joe’s uncle?

Neural network language models have achieved about 97% accuracy on a particular set of Winograd schemas. This roughly equals human performance.

In each sentence pair, the one-word difference can change which thing or person a pronoun refers to. Answering these questions correctly seems to require commonsense understanding. Winograd schemas are designed precisely to test this kind of understanding, alleviating the Turing test’s vulnerability to unreliable human judges or chatbot tricks. In particular, the authors designed a few hundred schemas that they believed were “Google-proof”: A machine shouldn’t be able to use a Google search (or anything like it) to answer the questions correctly.

These schemas were the subject of a competition held in 2016 in which the winning program was correct on only 58% of the sentences — hardly a better result than if it had guessed. Oren Etzioni, a leading AI researcher, quipped, “When AI can’t determine what ‘it’ refers to in a sentence, it’s hard to believe that it will take over the world.”

However, the ability of AI programs to solve Winograd schemas rose quickly due to the advent of large neural network language models. A 2020 paper from OpenAI reported that GPT-3 was correct on nearly 90% of the sentences in a benchmark set of Winograd schemas. Other language models have performed even better after training specifically on these tasks. At the time of this writing, neural network language models have achieved about 97% accuracy on a particular set of Winograd schemas that are part of an AI language-understanding competition known as SuperGLUE. This accuracy roughly equals human performance. Does this mean that neural network language models have attained humanlike understanding?

Not necessarily. Despite the creators’ best efforts, those Winograd schemas were not actually Google-proof. These challenges, like many other current tests of AI language understanding, sometimes permit shortcuts that allow neural networks to perform well without understanding. For example, consider the sentences “The sports car passed the mail truck because it was going faster” and “The sports car passed the mail truck because it was going slower.” A language model trained on a huge corpus of English sentences will have absorbed the correlation between “sports car” and “fast,” and between “mail truck” and “slow,” and so it can answer correctly based on those correlations alone rather than by drawing on any understanding. It turns out that many of the Winograd schemas in the SuperGLUE competition allow for these kinds of statistical correlations.

Rather than give up on the Winograd schemas as a test of understanding, a group of researchers from the Allen Institute for Artificial Intelligence decided instead to try to fix some of their problems. In 2019 they created WinoGrande, a much larger set of Winograd schemas. Instead of several hundred examples, WinoGrande contains a whopping 44,000 sentences. To obtain that many examples, the researchers turned to Amazon Mechanical Turk, a popular platform for crowdsourcing work. Each (human) worker was asked to write several sentence pairs, with some constraints to ensure that the collection would contain diverse topics, though now the sentences in each pair could differ by more than one word.

The researchers then attempted to eliminate sentences that could allow statistical shortcuts by applying a relatively unsophisticated AI method to each sentence and discarding any that were too easily solved. As expected, the remaining sentences presented a much harder challenge for machines than the original Winograd schema collection. While humans still scored very high, neural network language models that had matched human performance on the original set scored much lower on the WinoGrande set. This new challenge seemed to redeem Winograd schemas as a test for commonsense understanding — as long as the sentences were carefully screened to ensure that they were Google-proof.

However, another surprise was in store. In the almost two years since the WinoGrande collection was published, neural network language models have grown ever larger, and the larger they get, the better they seem to score on this new challenge. At the time of this writing, the current best programs — which have been trained on terabytes of text and then further trained on thousands of WinoGrande examples — get close to 90% correct (humans get about 94% correct). This increase in performance is due almost entirely to the increased size of the neural network langua

Have these ever larger networks finally attained humanlike commonsense understanding? Again, it’s not likely. The WinoGrande results come with some important caveats. For example, because the sentences relied on Amazon Mechanical Turk workers, the quality and coherence of the writing is quite uneven. Also, the “unsophisticated” AI method used to weed out “non-Google-proof” sentences may have been too unsophisticated to spot all possible statistical shortcuts available to a huge neural network, and it only applied to individual sentences, so some of the remaining sentences ended up losing their “twin.” One follow-up study showed that neural network language models tested on twin sentences only — and required to be correct on both — are much less accurate than humans, showing that the earlier 90% result is less significant than it seemed.

So, what to make of the Winograd saga? The main lesson is that it is often hard to determine from their performance on a given challenge if AI systems truly understand the language (or other data) that they process. We now know that neural networks often use statistical shortcuts — instead of actually demonstrating humanlike understanding — to obtain high performance on the Winograd schemas as well as many of the most popular “general language understanding” benchmarks.

The crux of the problem, in my view, is that understanding language requires understanding the world, and a machine exposed only to language cannot gain such an understanding. Consider what it means to understand “The sports car passed the mail truck because it was going slower.” You need to know what sports cars and mail trucks are, that cars can “pass” one another, and, at an even more basic level, that vehicles are objects that exist and interact in the world, driven by humans with their own aAll this is knowledge that we humans take for granted, but it’s not built into machines or likely to be explicitly written down in any of a language model’s training text. Some cognitive scientists have argued that humans rely on innate, pre-linguistic core knowledge of space, time and many other essential properties of the world in order to learn and understand language. If we want machines to similarly master human language, we will need to first endow them with the primordial principles humans are born with. And to assess machines’ understanding, we should start by assessing their grasp of these principles, which one might call “infant metaphysics.”

Training and evaluating machines for baby-level intelligence may seem like a giant step backward compared to the prodigious feats of AI systems like Watson and GPT-3. But if true and trustworthy understanding is the goal, this may be the only path to machines that can genuinely comprehend what “it” refers to in a sentence, and everything else that understanding “it” entailS.

 

More on the Zohar

 https://en.wikipedia.org/wiki/Zohar

Within Orthodox Judaism the traditional view that Shimon bar Yochai was the author is prevalent. R' Menachem Mendel Kasher in a 1958 article in the periodical Sinai argues against the claims of Gershom Scholem that the Zohar was written in the 13th Century by R' Moses de León.^[29] He writes:

1. Many statements in the works of the Rishonim (medieval commentors who preceded de León) refer to Medrashim that we are not aware of. He writes that these are in fact references to the Zohar. This has also been pointed out by R' David Luria in his work "Kadmus Sefer Ha'Zohar".
2. The Zohar's major opponent Elijah Delmedigo refers to the Zohar as having existed for "only" 300 years. Even he agrees that it was extant at the time of R' Moses de León.
3. He cites a document from R' Yitchok M' Acco who was sent by the Ramban to investigate the Zohar. The document brings witnesses that attest to the existence of the manuscript.
4. It is impossible to accept that R' Moses de León managed to forge a work within the scope of the Zohar (1700 pages) within a period of six years as Scholem claims.
5. A comparison between the Zohar and de León's other works show major stylistic differences. Although he made use of his manuscript of the Zohar, many ideas presented in his works contradict or ignore ideas mentioned in the Zohar. Luria also points this out.
6. Many of the Midrashic works achieved their final redaction in the Geonic period. Some of the anachronistic terminologies of the Zohar may date from that time.
7. Out of the thousands of words used in the Zohar, Scholem finds two anachronistic terms and nine cases of ungrammatical usage of words. This proves that the majority of the Zohar was written within the accepted time frame and only a small amount was added later (in the Geonic period as mentioned).
8. Some hard to understand terms may be attributed to acronyms or codes. He finds corollaries to such a practice in other ancient manuscripts.
9. The "borrowings" from medieval commentaries may be explained in a simple manner. It is not unheard of that a note written on the side of a text should on later copying be added to the main part of the text. The Talmud itself has Geonic additions from such a cause. Certainly, this would apply to the Zohar to which there did not exist other manuscripts to compare it with.
10. He cites an ancient manuscript that refers to a book Sod Gadol that seems to in fact be the Zohar.

 

THE AUHENTICITY OF THE ZOHAR

Some people are worried that the Zohar  has been proved  not to stem from R' Shimon bar Yochai. I very strongly recommend reading the series of posts starting here 

https://www.chabad.org/kabbalah/article_cdo/aid/380410/jewish/The-Zohars-Mysterious-Origins.htm?gclid=CjwKCAiA78aNBhAlEiwA7B76pxx6BsNATgWtrz5rhxU__H8BWMbLG8JeAQUK3ICh80UPRJ6MgvpsPBoCA7kQAvD_BwE

It seems to me that Rabbi Miller  has done an excellent job of refut9ngh these academic critics. I would appreciate comments.