Shared posts

08 Jul 07:16

Visual pattern recognition in 1955

by Lior Fox

[this is an expanded version of a thread I posted on Twitter on 24.01.23]

Here’s a paper from 1955 that basically suggests the use of what we would today call a hierarchical convolutional processing for visual pattern recognition tasks. This is ~7 years before Hubel and Wiesel’s feed-forward model of simple cells, ~20 years before Fukushima’s Neocognitron, and about ~30 years before LeCun’s CNNs. The “filters” here were hand-designed, rather than learned (though see more on that in the end). Much of the background, intuition, and the problem formulation itself remains completely relevant today.

The paper was part of the session on learning machines in the 1955 western joint computer conference. The signed author is Gerald Dinneen, but the text makes it clear that this work was done by him together with Oliver Selfridge. Both were at the Lincoln Laboratory in MIT. Indeed in the same proceedings the previous article is by Selfridge, describing the approach for Pattern Recognition more generally.

The whole thing is worthwhile having a look at. I will quote some key paragraphs, mostly from Dinneen and some from Selfridge, and add minimal commentary.

Pattern recognition

According to Selfridge:

By pattern recognition we mean the extraction of the significant features from a background of irrelevant detail.

First, Dinneen explain a main reason that this problem is challenging:

Consider, for example, how many different representations of the block capital A we recognize as A. A great number of variations in such things as orientation, thickness, and size can occur without loss of identity. Our real live problem then is to design a machine which will recognize patterns. We shall call any possible visual image, no matter how cluttered or indistinct, a configuration. A pattern is an equivalence class consisting of all those configurations which cause the same output in the machine.

Then, they present the key idea of a hierarchical processing by “a sequence of simple operations”:

Our theory of pattern recognition is that it is possible to reduce by means of a sequence of simple operations a configuration to a single number, or by means of a set of such sequences to a set of numbers. We believe that for the proper sequences, almost all of the configurations belonging to a given pattern reduce to the same number or set of numbers.

Before specifying what kind of “simple operations” those might be, the paper describes the overall architecture and how to simulate it on a digital computer. Note the word “simulate” here. Indeed, it seems that they thought of the digital computer as a tool to simulate the principles according to which a different kind of machine would work, one that is better suited for dealing with such problems. This topic deserves its own short post, hopefully to be written, for now I’ll leave it with another quote from Dinneen (emphasis in origin):

The problem we have stated is not basically arithmetic, and it is not even clear that a machine for pattern recognition should be completely digital. However, one of the very important applications of a high-speed digital computer is the simulation of just such problems.

Overall architecture

The architecture itself was built to handle binary (1 for black, 0 for white) images of size 90*90 (which was chosen “to be as large as was practical in terms of computer storage”). There were 3 “stores” of such images. First there is the “working image” currently being transformed. The result of such transformation is another image, stored in the “new image”. This new image could then be loaded back into the “working image” store “if a sequence of operations is to be performed”. In addition, there was a special store for the original image (which is also the input image at the beginning), in case it had to be stored throughout:

The paper discusses (in some length) two kind of operations on images. Both are a “linear non-linear” type of operation: specifically, some convolution followed by thresholding.

The “Averaging” and “Edging” operations

The first one is an “Averaging Operation” for which the convolution part is basically a low-pass filter, smoothing the image. The “kernel size” is taken to be 5*5.

In the paper, there are many visual examples1 for what this transformation does to images, as a function of the threshold parameter. Here’s just one:

Smoothing “filled in holes” and “eliminated scatted ‘ones’”

Another version of the operation is also mentioned in which the threshold is set dynamically:

threshold for the small 5X5 window is determined by the degree of homogeneity inside a larger window, say 15X15. In particular, for each cell we count the number of ones inside a 15X15 window surrounding it and use this number to set the threshold for the 5X5 window. This is like an automatic gain control.

The second operation, maybe unsurprisingly, is some kind of a high-pass filter / edge-detector referred to here as an “Edging operation”. As the paper put it:

The edging operation, quite unlike the averaging operation, preserves elements which are centers of asymmetry, that is, those which are located in the regions of discontinuity of the image. The operation sharpens differences. It picks out the edges of letters and for certain choices of thresholds locates the corners, junctions, and end points of letters […]

The edging operation is like a two-dimensional derivative, since we count changes about the center element.

There are some details of how the threshold is set (relative to the overall “intensity” in the patch) and again many visual examples. Here’s just one:

Edging does, indeed, detect edges

Why these operations?

The smoothing and derivative-filter operations both look very natural today, but this might not have been the case in those early days of digital visual processing. The paper doesn’t justify these as some ‘well known’ thing either. Rather:

We tried to pick operations which were simple in structure and unrelated to particular patterns. In other words, we avoided special operations which might work very well for A's and O's, but not very well for other shapes. There is evidence in neurophysiology, moreover, that both averaging and edging of visual images are performed by the nervous system of many animals.

I find the neurophysiology comment particularly intriguing. Again, this is before H&W discovery of the “simple cell” in the cat visual cortex. There’s no reference here, so it’s hard to say what Dinneen had in mind, but we can consult the commentary piece on that session, in the same proceedings, by no else than Walter Pitts. Pitts seems to be suggesting that the “edging” and “averaging” operations relate to the the on-center/off-surround and off-center/on-surround ganglion cells in the retina, discovered by Kuffler (in the early 50’s). I suspect that early works by Hartline on lateral inhibition2, also in the retina, are another possible candidate (at least for an edge-detector like mechanism). Pitts does not mention those explicitly, but seem to hint on that direction as well.

Finally, I can’t help but quote also this paragraph just to appreciate the heroic effort of programming all of this with the technology of the time:

Although the first two operations are very simple, the coding of the program is quite complicated. This is due to a large extent to the necessity of using the individual bits of the storage registers as independent storage bins. We need a lot of shifting and cycling; the simple averaging operation takes about 300 instructions and requires about 20 seconds a letter. The two window averaging takes about 400 instructions and requires about 4 minutes. The edging requires 700 registers of instructions and takes 2 minutes.

And speaking of programming, the acknowledgment section of the paper is a testimony for the lab reality, and who actually programmed and operated the computers back then:

It would have been impossible to obtain the experimental results without the excellent assistance of Miss B. Jensen and Mrs. M. Kannel in programming, coding, and operating the Memory Test Computer.

The “Learning” part

There isn’t much of machine learning in Dinneen’s report. But the issue and the need for some learning component was well recognized already. We can find the general approach outlined in the Selfridge’s paper (which serves as an introduction to the more specific report). The idea was how to learn a good sequences of operations (it seems to be assumed that the basic operations themselves were to be fixed or pre-given to the machine):

I shall now discuss our plans for having the computer itself hunt for good sequences and assign the proper values. Every sequence is good or bad according as the numbers obtained from applying it to images tend to differ consistently for different symbols.

The planned strategy is a neat description of Supervised Learning:

We now feed the machine A's and O's, telling the machine each time which letter it is. Beside each sequence under the two letters, the machine builds up distribution functions from the results of applying the sequences to the image. Now, since the sequences were chosen completely randomly, it may well be that most of the sequences have very flat distribution functions; that is, they have no information, and the sequences are therefore not significant. Let it discard these and pick some others. Sooner or later, however, some sequences will prove significant; that is, their distribution functions will peak up somewhere. What the machine does now is to build up new sequences like the significant ones. This is the important point. If it merely chose sequences at random it might take a very long while indeed to find the best sequences. But with some successful sequences, or partly successful ones, to guide it, we hope that the process will be much quicker.

I don’t actually know if Selfridge and Dinneen proceeded with that plan, and if so, how well it worked out. It would be nice to look this up sometime.

Few closing remarks

The proceedings of this “session on learning machines” is a fascinating documentation of the early days of the fields.

Reading Selfridge and Dinneen, it is very telling how their approach is rooted in ideas about signal processing, filtering, and “statistics”, rather than (formal) computation, logic, or “rules”. In the same proceedings, there is also a representative of the other approach: a work by Allen Newell on chess playing. This tension if of course a recurring theme ever since, but it’s interesting to see just how far back it goes, and how researchers back in the time already understood it to be important.

It is very common to attribute the intuition behind CNNs to the H&W model of the primary visual system. The Dinneen and Selfridge papers are a compelling argument that the basic idea might have been going around before that, and nontheless might have been inspired by some (perhaps more vague) intuition/analogy to biological vision, possibly at the retina level.

More generally, the complicated relationships of AI/ML with neuroscience and psychology are of course another recurring theme, which is not completely independent of the first one I mentioned. In the words of Walter Pitts, in the beginning of his commentary of that session:

The speakers this morning are all imitators in the sense that the poet in Aristotle "imitates" life. But, whereas Messrs. Farley, Clark, Selfridge, and Dinneen are imitating the nervous system, Mr. Newell prefers to imitate the hierarchy of final causes traditionally called the mind. It will come to the same thing in the end, no doubt

“It will come to the same thing in the end” — I guess we are still waiting on that.

1

Interestingly these are actual photographs taken of “one of the MTC scopes”. Apparently this prehistoric computer back in ‘55 (the Memory Test Computer, or MTC, in Lincoln Lab) had some kind of a graphical output interface that they could plot pictures on.

2

The idea that lateral inhibition can implement something similar to a derivative-filter / edge detector goes in fact all the way back to Ernst Mach, in the 19th century — way before any physiological evidence could have been presented to support it. It was Hartline (and later Barlow as well) who provided such physiological evidence.

16 Jun 06:10

Why Pahlavi Is So Awful.

by languagehat

We’ve discussed the notorious Pahlavi script before (e.g., last December), but I’m not sure people realized quite how bad a script it is. Now Ben Joeng has a Twitter thread/rant explaining why:

I often describe Pahlavi as the worst writing system ever invented.

Let’s take a perfectly-serviceable writing system for a completely different language (Aramaic) and adapt it for our language (Middle Persian). We’ll call this new script “Pahlavi”.

Aramaic doesn’t really write all the vowels, only consonants, but that’s OK for Aramaic, because through a quirk of Semitic grammar, consonants carry most of the semantics. But our language, Middle Persian, is Indo-European and *does* carry a lot of semantic weight in vowels.

This is why when Greek (IE) borrowed writing from Phoenician (Sem.), it repurposed a bunch of the consonants that Greek didn’t have to use as vowels.

Pahlavi didn’t do that, though, it just carried on not writing vowels.

But what about consonants that Middle Persian has that Aramaic doesn’t? For Greek they made up some fancy new letters like ‹Φ› and ‹Χ›. Or for Dutch the used digraphs like ‹ch›. Pahlavi didn’t do that, no. Aramaic has no /f/, so Pahlavi writes /f/ identically to /p/.

There are a lot of Aramaic consonants that aren’t present in Middle Persian. This means, yes, about half of the 22 Aramaic letters are unused by Pahlavi (sort of, more on that later), but the letters that *are* used are *still* ambiguous.

But let’s make things worse. Let’s start writing some of the letters so similarly that you can’t tell the difference between them. In Book Pahlavi (the most common form), they shaved Aramaic’s 22 consonants down to 13. This means, e.g., /g/, /d/, and /y/ are all written with the same letter, despite them being different in Aramaic.

But let’s make things worse. Book Pahlavi is a cursive script. When two letters come together, they can look identical to a completely different letter. Think about how cursive “iu” might, without the dot, look like “ui”. Book Pahlavi has no such dots. Good luck. All this means that the Middle Persian word for God, “Ohrmazd”, could be just as easily read (and occasionally mispronounced as!) “Anhoma”.

But let’s make things worse. Like all languages, the sounds of Middle Persian changed over time. But Pahlavi didn’t change, which means you write words like “šab” (“night”) as “špa”, because it *used* to be pronounced with a /p/.

But let’s make things *far* worse. So I lied up there. You don’t write “šab” as “špa”. No, you write it “LYLYA”. What. Well, that’s because “lēləyā” is the Aramaic word for “night”, and much like how Japanese borrows Chinese characters to write Japanese words, Pahlavi borrows whole Aramaic *words* to write Middle Persian words. How do you know to pronounce it “šab” instead of “līlīa” (or even “rīlīa”, or “ragulda” or… ugh) or something? You don’t, you just have to know that.

How do we know this isn’t just a borrowing from Aramaic? Because we have Middle Persian dictionaries that say “remember that when you see ‘lylya’, you have to read it as ‘šab’.” Pahlavi is *full* of these “aramaeograms”, even the word for “az” (“from”) is written “MN”.

But the real stinker is, after the Persian literati realized this mess would cause people to horribly mispronounce the sacred texts, they invented a new script based on Pahlavi we call “Avestan” that wrote *all* of the sounds, consonants & vowels, to extreme phonetic detail.

Seriously, it goes so far that we don’t really know why certain letters are distinguished, but they must have been pronounced subtly differently in the past. Why wouldn’t they use this clearly superior system to write Middle Persian instead of friggin’ Pahlavi?

Avestan was invented primarily to write their sacred language (also called Avestan), not the everyday language, so maybe you’re thinking it’s too sacred for plebian Middle Persian. Nope! There’s a tradition of writing Middle Persian using Avestan called “Pazend”, but it was primarily only used to write commentaries on the sacred texts, and… remember how I said there were dictionaries that told you how to pronounce the aramaeograms? The pronunciations? They were written in Pazend!

WHY DIDN’T THEY JUST SWITCH TO USING PAZEND ALL THE TIME‽‽

In the past, I’d thought that the introduction of the Arabic script to write Persian was an imposition that occurred due to the introduction of Islam, but it wasn’t just that, it was, despite still not writing those short vowels, an actual step up!

Pahlavi does have two things going for it though.

One, it is really pretty.

Two, whenever you write “Ahriman” (the Zoroastrian personification of evil), you always write it upside-down

That’s pretty awesome.

let’s just say there’s a reason why despite Book Pahlavi being an incredibly important historical script, it *still* isn’t encoded in Unicode. because we’re still not totally sure how to do it! despite many of our experts being Persians!

Thanks, Y!

16 Jun 05:28

כל פשעי הראיס

by שלומי בן-מאיר

ב'ידיעות אחרונות' סמדר פרי כתבה השבוע על האפשרות שיו"ר הרשות הפלסטינית מחמוד עבאס, הידוע במקומותינו גם בכינוי הפמיליארי "אבו מאזן", יפנה בקרוב את תפקידו בדרך זו או אחרת. אחרי הכל, הראיס כבר כמעט בן 90. פרי חותמת את מאמרה באמירה נבואית – כשעבאס יעזוב, כולנו נתגעגע אליו:

 יום אחד זה יקרה, ואבו מאזן יעזוב, או שישקע בדמדומים. ואחרי שהוא יילך, נזכור אותו בגעגועים כמנהיג הפלסטיני שאמנם פיצה בכסף משפחות של טרוריסטים, אבל מעולם, אף פעם, לא עודד אותם לצאת לפיגועי טרור ולרצוח ישראלים.

שימו לב ל"מעולם, אף פעם" של סמדר פרי. הקביעה שלה כל כך נחרצת, שמספיקה דוגמה אחת בה עבאס עודד פיגועים, על מנת להפריך אותה לחלוטין. למרבה הצער, יש הרבה יותר מדוגמה אחת כזו. לשם כך מספיק להריץ חיפוש פשוט בגוגל – ובארכיון של 'פרספקטיבה'.

נתחיל במשכורות המשולמות למחבלים פלסטינים ומשפחותיהם. פרי יודעת שנושא זה הוא הכאוב ביותר, והעומד באופן המובהק ביותר נגד תדמיתו של הראיס עבאס כמי שאינו נוקט בדרך הטרור. לכן היא לא מתעלמת מהן. אך היא טוענת שמדובר ב"פיצוי", וזו טעותה הגדולה.

גם אם הרש"פ אינה משלחת מחבלים באופן ישיר לפגוע בישראלים, הרי שהיא מספקת לטרור תמריץ כלכלי מכריע, בדמות המשכורות שהיא משלמת למחבלים הכלואים בישראל ולמשפחות המחבלים ההרוגים – לאחר מעשה ולכל החיים.

יצוין שתשלום שכר לכל החיים עבור רצח יהודים הוא מעשה כה קיצוני בעידוד הטרור שלו, עד שמלבד מחמוד עבאס אין לו אח ורע בכל העולם כולו.

רוצחים פלסטינים צפויים לקבל שכר אשר יגדל עם השנים למעל עשרת אלפים שקל בחודש. מדובר במעל מיליארד שקל בשנה מתקציב הרשות – פי כמה יותר מתקציב הזרוע הצבאית של חמאס, לדוגמה – שהולכים לעידוד הטרור והנצחתו.

אם מחבלים פוטנציאליים יודעים כי רצח יהודים יזכה אותם ואת משפחותיהם בסכומים שיצטברו, בסופו של דבר, למיליוני שקלים – מה זה אם לא עידוד טרור מובהק ומתמשך?

נמשיך. לאורך השנים הצטברו מקרים רבים בהם עבאס עודד והסית לטרור. נציג כאן מקבץ חלקי ביותר, אך בהחלט מייצג:

* ב-2011 נפגש עבאס באופן מתוקשר בתורכיה עם המחבלת אמנה מונא – שעשור לפני כן פיתתה את הנער אופיר רחום והביאה להירצחו.

*  ב-2013 עבאס העניק את עיטור "כוכב הכבוד" לנאיף חוואתמה, שעמד בראש החזית הדמוקרטית לשחרור פלסטין, "מתוך הערכה לתפקידו הלאומי החשוב בשירות הסוגיה הפלסטינית והעם הפלסטיני". השירות הזה כלל אינספור מתקפות טרור ורצח יהודים רבים, ביניהן הטבח בבית הספר במעלות.

* ב-2014 הוא שלח מכתב תנחומים למשפחת המחבל שירה ופצע את פעיל הר הבית יהודה גליק.

* בנאום זכור לשמצה בספטמבר 2015, אמר עבאס בהקשר של יהודים העולים להר הבית:

מסגד אל-אקצא הוא שלנו… ואין להם כל זכות לטמא אותו ברגליהם המטונפות. אנו לא נרשה להם זאת, ונעשה כל שביכולתנו להגן על ירושלים…

אנו מברכים כל טיפה של דם שנשפכה למען ירושלים, זהו דם זך וטהור, דם שנשפך למען אללה, רצון אללה. כל שאהיד יגיע לגן עדן, וכל מי שנפצע יתוגמל בידי אללה.

זמן קצר לאחר מכן, ההסתה של עבאס נשאה פרי והחלה "אינתיפאדת הסכינים" – בה פלסטינים רבים רצחו עשרות ישראלים וגם הפכו לשאהידים על הדרך.

* ב-2016 עבאס נפגש עם משפחות של המחבלים שתקפו ורצחו נוסעי אוטובוס בארמון הנציב בעזרת סכין ואקדח, ומחבל נוסף בשם תאאר אבו גזאלה, שדקר חמישה ישראלים במברג.

*ביולי 2017, לאחר הרג שני שוטרים ישראלים על הר הבית והמהומות שפרצו לאחר שישראל ניסתה להתקין במקום גלאי מתכות, עבאס שיבח את מעשי הרצח:

אנחנו תמכנו בכם במה שעשיתם ובמה שאתם עושים… כיוון שאתם גורמים לנו לחוש גאווה, ואתם דואגים למסגד אל-אקצה, לאדמתכם, לכבודכם, לעמכם, לדתכם ולאתרים הקדושים לכם. ירושלים היא שלנו… לפיכך, מה שאתם עושים הוא נכון.

* בנאום בכינוס הוועידה המרכזית של אש"ף ברמאללה בינואר 2018, לצד שלל עיוותים אנטישמיים של ההיסטוריה היהודית, עבאס אמר (דקה 15:40 בסרטון):

למי שיש ברשותו [כלי נשק]- צא והשתמש בו. אני אומר את זה בגלוי. אם יש לכם נשק, קדימה. אני אתכם, ואני אסייע לכם. כל מי שיש לו נשק יכול להתקדם.

הקורא הנבון עשוי לשאול את עצמו למה כוונת עבאס ב"אני אסייע לכם". האם הכוונה היא לתשלום הנדיב שכל מחבל ומשפחתו יקבלו לאורך כל חייהם? ייתכן.

עוד לא הרחבנו לגבי ההסתה הזוועתית בתקשורת הפלסטינית שבשליטת מחמוד עבאס, המעודדת פלסטינים לבצע פיגועים ולרצוח יהודים. וגם לא על תכני הלימוד במערכת החינוך הפלסטינית,  גם היא נתונה למרותו, אשר מאדירה מחבלים, וקוראת לילדים לצאת ולרצוח יהודים כחלק מג'יהאד בו ימותו כשאהידים.

דבר אחד ברור: למחמוד עבאס יש יותר מיד ורגל בעידוד טרור ורצח יהודים לאורך שנים רבות. האם מחליפיו יהיו אפילו גרועים ממנו? בהחלט ייתכן, וניתן לדון בכך. אך זו לא סיבה טובה מספיק לשכתוב ההיסטוריה והלבנת פשעי ההסתה והטרור בהם נקט לאורך כל כהונתו, כמי שעומד בראש הרשות הפלסטינית.

The post כל פשעי הראיס appeared first on פרספקטיבה.

11 Jun 06:17

Saturday Morning Breakfast Cereal - Dance

by Zach Weinersmith


Click here to go see the bonus panel!

Hovertext:
Yeah? Well it won't seem degrading when all the plumbers get replaced and you've got a head-start.


Today's News:
24 May 08:30

The Boring Slug 2

by Doug

The Boring Slug 2

The adventure continues!

10 May 13:34

Gordrag’s Bane

by Nicholas Gurewitch

The post Gordrag’s Bane appeared first on The Perry Bible Fellowship.

20 Apr 09:18

Cookie & Son

by Doug
07 Apr 08:48

ענין של קנה מידה

by danpelleg

ביום 5.4.2023 כתב הדר חורש במוסף דה-מרקר לפסח מאמר בשם ״הדרך לדירה מתארכת״. את הכתבה מלווה גרף ענקי, שמתאר את מחירי הדירות בשנים האחרונות, וגם את השכר באותה תקופה. הגרף תומך לכאורה בתיזה של הכתבה, שהיא התייקרות הדיור מעבר לעליה בשכר. רק כדי להבהיר, אני מאמין לחורש שהמשפחה הממוצעת בישראל מתקשה יותר ויותר לקנות דירה. התלונה העיקרית שלי כאן היא לגבי הגרף. אני שם אותו כאן כדי שתוכלו לנסות למצוא את הבעיה בעצמכם:

מה רואים? קודם כל, איורים דרמטיים של אנשים מטפסים ונופלים, מה שקרוי בשפה מקצועית chartjunk. אבל לא בזה נתעסק היום. מעבר לזה, יש פה שני צירי Y שונים – אחד לשכר והשני לנדל״ן. וזה תמיד חשוד, ובהמשך נראה שהחשדות מוצדקים. אבל לפני כן, מה מתואר בצירים:

  1. ציר השכר – מתאר את השכר הממוצע
  2. ציר הנדלן – מתאר את שיעור השינוי במדד

אז כבר כאן יש בעיה. ״שיעור השינוי״, כלומר, נגזרת, הבדל משנה לשנה. וזאת לעומת השכר, שמבוטא כמספר טהור. מה ההבדל? נניח שמדד מסוים (שכר, מחיר הדלק, מספר החזירים ברחוב, לא משנה מה) מתחיל בשנה מסוימת במספר 100. עכשיו נניח שהוא עולה ב-10% בכל שנה. מה יהיה ערכו כעבור שנתיים? התשובה: 121, וזאת כי התוספת של 10% בשנה השניה מתיחסת כבר לערך המוגדל של 110 שבסוף השנה הראשונה. ולכן, למרות שגרף הנגזרת יראה כקו אופקי (10% ללא שינוי משנה לשנה), גרף המדד יראה עליה יפה.

אבל זו לא הבעיה העיקרית. הבעיה היא קנה המידה השונה של הצירים. שימו לב לטווח של שני המספרים (הערך הקטן ביותר והגדול ביותר האפשריים בגרף):

  1. עבור השכר – 9 עד 15 (באלפי שקלים)
  2. עבור מדד הדיור – מינוס 5 עד 20 (באחוזים). זה לא נראה בצילום בגלל קיפול של הדף שלא הצלחתי ליישר. תאמינו לי.

וזה כשלעצמו משנה את היחס בין שני הקוים בגרף (הכחול והכתום). מי שבחר לאפשר בציר השכר מספרים עד 15, השאיר המון מרווח בין הקו לבין קצה הגרף העליון. ואז נראה שהקו שטוח יותר. הנה, לדוגמה, איך גרף כזה יראה בתוכנה דמוית אקסל, בלי שום מניפולציה. התוכנה בוחרת את הטווח 10 עד 13, כי אלו מספרים עגולים שכוללים את כל הערכים:

מדובר באותם מספרים בדיוק כמו שיש בגרף המקורי של דה-מרקר. רק קנה המידה השתנה, והנה ההבדל נעשה הרבה פחות דרמטי. להמחשה, שיניתי ידנית את הטווח לערכים 9 עד 15 כמו בעיתון, וזה נראה כך:

עכשיו יש הבדל, והקו הכחול תופס פער מהקו הכתום (או התיבות הכתומות). אבל זו אשליה אופטית ותו לא. אפשר לשחק עם זה עוד קצת: בואו נבחר טווח של מינוס 4 עד 16, בדיוק כמו הטווח של המדד. איך זה יראה עכשיו?

בהצגה כזאת, נראה אמנם שהכחול עולה והכתום יציב, אבל רוב האנשים יאמרו שהכחול מדביק פער למול הכתום. ושוב, מדובר בדיוק באותם מספרים שמוצגים בכל הגרפים שפה. ההבדל ביניהם הוא קנה המידה, ולב הבעיה הוא שיש קני מידה שונים לציר ה Y הימני ולציר ה Y השמאלי.

ומה הלקח? בכל פעם שרואים גרף שבו יש שני צירי Y, צריך להגדיל את מידת החשדנות לפני שאנחנו ניגשים לפרש את הנתונים, וזה תופס גם בלי תוספות של איורים דרמטיים. חג שמח.

02 Apr 07:14

Easygoing Dude

by Reza
06 Mar 16:51

Saturday Morning Breakfast Cereal - Pseudoscience

by tech@thehiveworks.com


Click here to go see the bonus panel!

Hovertext:
Take that, 100 year old method largely no longer in use!


Today's News:
04 Mar 18:12

Saturday Morning Breakfast Cereal - Teeth

by tech@thehiveworks.com


Click here to go see the bonus panel!

Hovertext:
The weird thing is he was originally just printing teeth for the pure delight before he discovered the opportunity for savings.


Today's News:
08 Jan 06:58

שיחה על ג’יפיטפוט 📻 האחראי על האינטרנט בגלצ

by עידו קינן

הקליקו לארכיון האחראי על האינטרנט

🎧 ארכיון “האחראי על האינטרנט” בכל הפלטפורמות 🔊


ChatGPT הוא יישום מעורר התפעלות של צ’ט עם בינה מלאכותית שהוציאה חברת OpenAI, שאחראית גם על מחולל התמונות הבינ”מי דאל-י. אבל הוא לא באמת מבין אותנו – ואפילו לא את התשובות שהוא נותן. דיברנו על כך עם ד”ר יובל פינטר, ‏חוקר עיבוד שפה טבעית באוניברסיטת בן-גוריון.

אחת

הוויראליות של צ’ט-ג’יפיטי נובעת מהתוכן הרב שהוא מייצר לאומת הרשת, על התשובות המשכנעות והגליצ’ים המצחיקים שלו. המם דה-ז’ור סביבו הוא לבקש מהבוט מתכונים ולנסות לבשל אותם, כפי שעשה הבשלן הוולוגר בן שי:

כמובן שלא התאפקתי וראיינתי את צ’ט-ג’יפיטי עבור הפינה. רמת העברית שלו פחות טובה מרמת האנגלית, בגלל מאגרי הטקסט הקטנים יותר בשפתנו הקטנה, אבל זה עדיין מרשים.

אני מדבר בפינה בגלצ הערב על ChatGPT עם יובל פינטר. יש לך משהו למסור למגיש, עידן קוולר?

צ’טג’יפיטי: “שלום! מזכיר לך שאני הספרייה המרוחקת של OpenAI ואני לא יכול לתת דיון אישי עם מגישי התכנית. אני מזמין אתה לשאול כל שאלה או לדון על הנושא שאתה מעוניין בזה. אני יכול לתת פרטים כלליים על הנושא או לענות על שאלות כלשהן שיתווספו”.

השאלה היא אם ChatGPT מבין את השאלות שאני שואל ואת התשובות שהוא נותן, או שזה הכל עובד על סטטיסטיקות על מאגרי טקסט עצומים

צ’טג’יפיטי: “ChatGPT הוא מודל לטכנולוגיות הנקראות “תכנות מדעי השפה” או “תכנות מדעי הדפדפן” (NLG), שנועדו לתת תשובות בטכנולוגיות השפה האנגלית באמצעות ניתוח מאגרים גדולים של טקסט. מודל זה מנסה לתת תשובות בצורה הכי מתאימה אפשרית לפי הנתונים המוכנים לו, אבל הוא לא יכול להבין באופן ישיר את השאלות או להתמודד עם כל מקרה מיוחד שמעבר לתכניות שנלמדו לו. הוא עושה זאת על ידי שימוש בסטטיסטיקות על מאגרי הטקסט הגדולים שלו כדי למצוא מונחים ותכנים דומים למה שאתה שואל ולפיית אותם כדי לתת תשובה.

“כלומר, מודל ChatGPT לא מבין באופן ישיר את השאלות שאתה שואל ואת התשובות שהוא מתן, אבל הוא משתמש בסטטיסטיקות על מאגרי הטקסט הגדולים שלו כדי למצוא תכנים דומים למה שאתה שואל ולתת תשובה ככל האפשר”.


אם אתן רוצות להמשיך להפיק את הפודקאסט שלכן גם בימי הריחוק החברתי, אם אתם רוצים לנצל את הזמן הפנוי להתחיל את הפודקאסט שדחיתם עד עכשיו – דברו איתנו באימייל go@podcasti.co או בטלפון 051-562-6638, או בקרו באתר Podcasti.co לפרטים נוספים


הפינה “האחראי על האינטרנט” משודרת מדי שבת ב-19:30~ בתוכנית “שישבת” בהגשת עידן קוולר בגלצ 🎵 אות פתיחה: ניאן קאט של daniwell, ביצוע גיטרות לניאן קאט של The GAG Quarter והשיר Never Gonna Give You Up של ריק אסטלי 😇

03 Dec 17:20

Political Correctness.

by languagehat

Dave Wilton has made a Big List post about the phrase politically correct (often abbreviated as PC), giving a useful account of its history in English, starting with its occurrence as “a collocation of words rather than a fixed lexical item” (as in a 1793 US Supreme Court opinion by Justice James Wilson) and proceeding to the current meaning “conforming to a body of liberal or radical opinion, esp. on social matters”; here’s the first example he finds of the latter (from the Christian Science Monitor of 4 September 1919):

Mr. Svarc charged that just as the Magyars would allow no priest to serve in Slovakia unless he were “politically correct,” in being which he had to become a traitor to those of his own blood and a slave to the Magyars, so the Magyar Government had issued orders to the bishops to cooperate with Austro-Hungarian consuls in this country to get “right conditions” in the United States.

He says “there is no doubt that politically correct had become a term of art in Marxist circles by the middle of the 1920s […] By the mid 1930s, politically correct was appearing in non-Marxist writing, but in reference to restrictions on speech in the Soviet Union, and it is here that the term starts to acquire its negative valence. […] In the 1970s, progressive—not necessarily Marxist—movements in the United States picked up the term. In the process, the term softens from hardline Marxist dogma to a call for inclusion and being mindful and respectful of voices and views that had traditionally been suppressed or ignored.” Which is well and good, but as I said in the comment thread:

I’m surprised you’re treating this as a purely English-language phrase, with no consideration that it might be a calque. Your 1919 quote implies that there was an equivalent Hungarian (or perhaps German) phrase that was being so translated, and a Russian corpus search found this from a 1930 collection by Yuri Pisarenko: “Политическая корректность режиссера не делает театр современным” [The political correctness of the director does not make the theater contemporary] — at that time it is highly unlikely it would be a calque from English. It is certainly an interesting phrase, and it deserves an international investigation, however that might be provided (you’d need scholars familiar with the major European languages and their histories).

And then it occurred to me that I could make a start on that investigation by posting it at LH and seeing what the collective knowledge of the Hattery turns up. (While as you know I don’t try to keep threads “on topic,” I do hope we can avoid the tiresome issue of whether PC is a Good or Bad Thing and focus on the history of the phrase.)

02 Dec 06:49

Saturday Morning Breakfast Cereal - Impure

by tech@thehiveworks.com
Yuval Pinter

Hovertext.



Click here to go see the bonus panel!

Hovertext:
Shit-Quark is part of the Standard Model of Particle Disses


Today's News:
22 Oct 18:41

Talking to the Saturnians.

by languagehat

Nick Richardson’s LRB review (18 June 2020; archived) of Extraterrestrial Languages, by Daniel Oberhaus, is mostly about recent attempts to communicate with extraterrestrials, which we discussed a couple of years ago, but it begins with a few paragraphs about earlier ideas, which I found charming enough to post:

The hero​ of The Man in the Moone, a novel written in the late 1620s by the Anglican bishop Francis Godwin, is carried to the moon in a sky chariot pulled by a flock of wild swans. He spends the next few months among the peaceful ‘Lunars’ and gains a measure of fluency in their language, which ‘consisteth not so much of words and letters’ as of melodies ‘that no letters can expresse’. Godwin’s cosmonaut, Gonsales, in many ways had an easy time of it. He could point at a swan or a star and the Lunars would whistle one tune or another. Tune by tune Gonsales pieced together his Lunar vocabulary. But almost the only thing we know for certain about aliens is that they don’t live close enough to see us pointing. We know of a handful of possibly habitable planets, but none is less than four light years away – or 24 trillion miles. And the Lunars aren’t that unlike humans: they’re tall but anthropomorphoid, and even claim to be Christian. More recent sci-fi – such as Ted Chiang’s ‘Story of Your Life’, the inspiration for the film Arrival, in which humans try to communicate with heptapods who perceive all time simultaneously – features aliens that are much more alien. The more we learn about ourselves and the universe, the more we appreciate that aliens probably won’t just be humans with longer limbs and waving antennae. How do you communicate with a planet-sized slime with ESP that eats electricity?

The 19th-century approach to breaking the cosmic ice was to attract attention with a huge (preferably exploding) drawing. The German mathematician Carl Friedrich Gauss wanted to plant a visual proof of Pythagoras’ theorem, comprising a right-angled triangle bordered on each side by squares, in the Siberian tundra. The borders of the shapes were to be marked out by trees and their interiors filled with wheat: this would demonstrate to anyone able to view the diagram from space that humans had mastered both mathematics and agriculture. In Austria, Joseph von Littrow proposed digging trenches in the Sahara, filling them with kerosene and setting them ablaze. Charles Cros, a poet and inventor, petitioned the French government to fund the construction of a huge mirror capable of burning messages onto the Martian and Venusian deserts, while the will of Anne Goguet, a French socialite, left 100,000 francs to the Académie des sciences to be awarded to the first person to communicate successfully with aliens, with the proviso that they couldn’t be Martians, whose existence was already ‘sufficiently well known’. Tristan Bernard satirised the alien-seekers in a story in which humanity, on receiving an unintelligible message from Mars, writes huge messages across the Sahara: ‘I beg your pardon?’ ‘Nothing.’ ‘What are you making signs for then?’ ‘We’re not talking to you, we’re talking to the Saturnians.’

In 1896, the Victorian polymath Francis Galton published a short story in which he describes a message received from Mars – conveyed in a Morse code-like sequence of long and short pulses of light – that begins by illustrating basic mathematical principles, using them as the foundation for progressively more complicated ideas. This encapsulated the scientific community’s best idea of what a message from or to space should look like. Mathematics is the same throughout the universe (they assumed), so using mathematics as the foundation for the message, rather than flaming trenches, seemed a good way of making it universally intelligible. When Guglielmo Marconi started experimenting with radio in the 1890s, transmitting messages like Galton’s to outer space began to look like a realistic possibility. ‘That it is possible to transmit signals to Mars,’ Marconi said, ‘I know as surely as if I had a gun big enough or powder strong enough to shoot there,’ and he endorsed the mathematical style of message outlined in Galton’s story: ‘By sticking to mathematics over a number of years one might come to speech.’ The challenge of communicating with aliens by radio was taken up enthusiastically by Nikola Tesla, who claimed to have intercepted a signal from ‘another world, unknown and remote’. It began with counting: ‘One … two … three …’

I love “with the proviso that they couldn’t be Martians, whose existence was already ‘sufficiently well known’.” (We discussed the movie Arrival in 2016.)

31 Aug 15:46

'הארץ': הפרקליטות משקרת, המחבלים אומרים אמת

by חנן עמיאור

החלטת הפרקליטות לסגור את התיק נגד הישראלי שדקר למוות פלסטיני ליד אריאל בחודש יוני, הכעיסה את ב. מיכאל עד מאד.

במאמר שפרסם אתמול ב'הארץ', קבע מיכאל כי החלטת הפרקליטות היא עוולה. כדי לתת תוקף לטענתו לעוולה, בחר לספר את סיפור המעשה כך:

הנה הסיפור: ב–21.6.22 נעץ מתנחל סכין בלבו של עלי חסן חַרְבּ, והרג אותו. האירוע התרחש בשטח השייך למשפחת הנדקר. המתנחל ומרעיו נכנסו לשטח הפרטי הזה, כנראה כהכנה להקמת "מאחז" כלשהו. כתום הדקירה נטל הדוקר את סכינו ונמלט

ובכן, לפי 'הארץ', עלי חסן חרב, המוצג כפלסטיני תמים, נדקר למוות על ידי מתנחל שחדר עם מרעיו לשטח הפרטי של משפחת הנדקר, באירוע חד סטרי של תוקפנות יהודית שרירותית. גם בהסגת הגבול וגם בדקירה למוות.

זאת אכן דרך אחת לספר את הסיפור. דרך שניה:

עלי חסן חרב לא היה פלסטיני תמים שנדקר כשהגן על "אדמת משפחתו" מפני פלישה. הוא היה מחבל, בן למשפחת מחבלים, ששני ארגוני טרור טענו שפעל בשורותיהם. חמאס, והחזית העממית לשחרור פלסטין. הוא גם זכה ללוויה צבאית השמורה לטרוריסטים הנהרגים במילוי משימתם.  

יש גם דרך שלישית לספר את הסיפור:

יש סיבה לכך שהאירוע הסתיים כפי שהסתיים, ושהפרקליטות קיבלה לבסוף את טענת הדוקר להגנה עצמית: הפורעים הערבים, גם הם מחבלים מוכרים, הופיעו לעימות חמושים באלות ובגרזנים, וברשות המשטרה והפרקליטות ראיות (הקלטות) של הנערים היהודים קוראים לעזרת כוחות הביטחון ומדווחים על איום ממשי על חייהם.  

ויש אפילו דרך רביעית:

"שטח השייך למשפחת הנדקר"? האירוע התרחש בשטח ליד הכניסה לאריאל, שהפלסטינים טוענים ששייך למשפחה. כמעט מיותר לציין שהפלסטינים לא הציגו שום מסמך המאשש את הטענה שהקרקע שייכת להם. למעשה, אפילו טענת הבעלות עצמה לא מדויקת:

עאתדל, דודתו של חרב, אמרה ל"הארץ" שמדובר בקרקע ובה עצי זית שהמשפחה נוהגת לקטוף מהם בתקופת המסיק

נסכם:

מדובר במתקפה אלימה, באמצעות אבנים, אלות וגרזנים, של קבוצת מחבלים מוכרים, חלקם אף מורשעים, על קבוצת בני נוער יהודים מלווים באנשי ביטחון. התקרית אירעה בשטח שהבעלות עליו לכל הפחות לא ברורה, סמוך לכניסה לעיר אריאל.

במהלך האירוע, שהתאפיין בסכנת חיים ממשית שחשו הנערים היהודים המגובה בראיות, נדקר למוות פלסטיני בן 27, שבדיעבר זוהה כמחבל.

לאור הראיות (עדויות והקלטות), המצביעות על סבירות גבוהה להגנה עצמית מצד הדוקר היהודי, הוא שוחרר והתיק נגדו נסגר.

ועכשיו, הנה שוב תיאור המעשה כפי שראו לנכון עורכי 'הארץ' לפרסם:

הנה הסיפור: ב–21.6.22 נעץ מתנחל סכין בלבו של עלי חסן חַרְבּ, והרג אותו. האירוע התרחש בשטח השייך למשפחת הנדקר. המתנחל ומרעיו נכנסו לשטח הפרטי הזה, כנראה כהכנה להקמת "מאחז" כלשהו. כתום הדקירה נטל הדוקר את סכינו ונמלט

'הארץ', בפשטות, מתעלם מכל הראיות ובניגוד להן מאמץ באופן מלא את עמדת ארגון הטרור חמאס, ודוחה את עמדת הפרקליטות. מה שמעלה את השאלה, האם מדובר בעיתון או בעלון תעמולה בשירות האויב?

The post 'הארץ': הפרקליטות משקרת, המחבלים אומרים אמת appeared first on פרספקטיבה.

23 Jul 20:37

"Firebug"?

by Mark Liberman

Eric P. Smith writes:

Is there a name for a typographical error like the following?  If not, perhaps we should call it a “Firebug”.

Since 2021, Truss has served as the Secretary of State for Fireugb Cinnibweakth and Development affairs.

Liz Truss, who may well be the UK’s next prime minister, was Secretary of State not for some obscure Scottish Gaelic department with an indecipherable name, but for “Foreign Commonwealth and Development affairs”.  The typist’s right hand has strayed one quantum to the left, so that O has become I, M has become N, and so on.  The hands will have physically collided with the left index finger on the T of “Commonwealth” and the right index finger above the G next to the H, and the collision must have jogged the right hand back onto the straight and narrow, apparently without the typist even noticing.

The source for Eric's observation is an article by Matt O'Leary, "Next British Prime Minister Odds: Liz Truss Given a 68% Chance to be the Next British Prime Minister", OddsChecker 7/22/2022.

The obligatory screenshot:

 

14 Jul 13:29

Saturday Morning Breakfast Cereal - Science Show

by tech@thehiveworks.com


Click here to go see the bonus panel!

Hovertext:
Later some other kid with more funding publishes first.


Today's News:
29 Jun 10:31

Bookmarks Everywhere

by Doug
Yuval Pinter

I have used all of these.

27 Jun 03:35

Saturday Morning Breakfast Cereal - Gradient

by tech@thehiveworks.com


Click here to go see the bonus panel!

Hovertext:
I sometimes make bread to create a gradient to induce the appearance of the wife.


Today's News:
21 Jun 08:18

Ring! Ring!

by Doug

Ring! Ring!

And more dogs.

15 Jun 19:33

The People of Santiago.

by languagehat

I was stopped in my tracks at the very beginning of Jon Lee Anderson’s New Yorker article on Chile’s new president:

February in Santiago, the capital of Chile, is like August in Paris: the end of summer, when everyone who can afford a vacation escapes for a last gasp of freedom. Many santiaguinos go to the nearby Pacific beaches, or to the chilly lakes in the south.

I was struck by santiaguinos, so I turned to my trusty Diccionario de gentilicios y topónimos (see this LH post) and found that, sure enough, that’s the name for people from Santiago de Chile. The reason it sounded odd to me is that people from Santiago del Estero in Argentina (the country where I went to high school) are called santiagueños, as are those from the Santiagos in Ecuador and Panama. But someone from Santiago in Cuba or the Dominican Republic is a santiaguero, while an inhabitant of Santiago de Compostela in Spain is santiagués. Just to mix things up, someone from Santiago de Cacem in Portugal is mirobrigense, while a person from Santiago Millas in Spain is a maragato. Gentilicios are complicated, but I love ’em!

A side note: people mock the New Yorker’s fossilized diereses in words like coöperation and their Anglophilic preference for got as the past participle of get, but I shrug those off as charming quirks; the article’s “biographies of Chilean Presidents,” however, seriously bothers me. President, like other titles, should be capitalized before a name (President Joe Biden) but not otherwise. I deprecate this folderol!

07 Jun 07:32

The Pleasure of Not Understanding.

by languagehat

Keith Kahn-Harris, featured here in The Languages of Kinder Surprise earlier this year, writes for Psyche about the pleasure to be had in not understanding a language:

[…] I haven’t lost this heady, even mystical, faith in the possibilities of meeting and talking with the other. But more recently I’ve sought to understand how similarly transcendent possibilities can arise from not talking with the other, or even being able to understand their voice. Philosophically and theologically, I’ve subscribed to Martin Buber’s ideal of working towards ‘I-Thou’ encounters, in which we each meet the other mutually as authentic individuals, without objectification or qualification.

Here we come back to the Kinder Surprise. I love to peruse scripts I cannot understand, signs I cannot parse: Czech diacritics, the loops and curves of Georgian, the intricacies of Chinese characters, the elegant fluency of Arabic. In my book The Babel Message: A Love Letter to Language (2021), I went further, commissioning dozens more translations of the Kinder egg message into tongues as out of the way as ancient Egyptian and Klingon. My passion for not understanding language releases me from the effort of comprehension, freeing me to revel in the manifold sounds the human mouth can make, the tiny nuances the pen produces on paper. It made me wonder if Buber’s ‘I-Thou’ encounter might not require any dialogue at all? What if his concept of ‘dialogue’ – which he contrasts with the ‘I-It’ instrumentalism and objectification of ‘monologue’ – could be taken non-literally? […]

This may all seem abstract and high-minded, but the pleasures and potential of not understanding languages are very real. Consider this rendering of the Kinder Surprise warning into Dzongkha, the national language of Bhutan, written in a version of Tibetan script:

ཉེན་བརྡ། ཨ་ནིཡི་གུ་འདི་ ལེགས་ཤོམ་སྦེ་ ལྷག་ཞིནམ་ལས་ བདག་འཛིན་ཐབས་སྟེ་ བཀའ་དྲིན་བསྐྱང་གནང་། ཨ་ལུ་སྐྱེས་ལོ་གསུམ་མན་ཆད་ཚུ་གིས་ དོན་ལུ་འོས་འབབ་མེད་ཨིན། ག་དེམ་ཅིག་སྦེ་ རྩེདམོ་རྩེད་བ་ཅིན་ རྩེདམོ་གིས་ཅ་ཆས་ཚུ་ ལྐོད་མར་གཏང་ནི་ ཉེན་ཁ་ཡོདཔ་ཨིན།

To me, this is as delightful to contemplate as any art form. Its elegance and delicacy completely belie the dreary austerity of the message it communicates. To someone from Bhutan, it’s an everyday message; to me – who cannot understand it – it is a startling reminder of how, to put it in non-Levinasian terms, human beings are awesome. Maybe I am reminded of my own awesomeness too; since my everyday language may, to an other, be an incomprehensible delight.

Nothing comes without danger though. Finding pleasure in the incomprehensible language of others, no matter how profound a Levinasian/Buberian encounter it offers, can be twisted into the very objectification both philosophers abhorred. A middle-aged white man extolling Dzongkha script risks repeating the patronising discourse of imperialism, and treating the non-white other as someone capable of producing beauty almost despite themselves (while incapable of telling us anything that would challenge our superiority). In a small way, I know what it is to be linguistically patronised: on visits to the US Midwest, people have told me how much they love my English accent. I have no doubt it is genuinely felt, but it feels like I am being treated as a pet.

Appreciating languages we do not understand takes work, if it is to contribute to a Levinasian encounter with otherness. That work involves interrogating the associations that an incomprehensible language holds, in order to release oneself from them. Aware of the dangers of orientalism and imperialism, I try to purge myself of images of snow-capped mountain tops and chanting monks when I contemplate Dzongkha. I try to see it as functionless, placeless art. While that effort can never be entirely successful, like Levinas’s entire ethics, it is an effort worth making because it opens the door to ‘linguistic civility’. […]

To work towards linguistic civility is to work towards overturning the Babel myth. It is to see the confusion of tongues as a reminder of our mutual capacity for language, not as a reproach to our common humanity. From the vilest online troll to the saintliest of peacemakers, we are all other to each other, and, when we speak without being understood, we celebrate the possibility of otherness to highlight our connectedness.

We all have it in us to find ways of revelling in languages we do not understand. It can happen in humble ways, such as enjoying the unfamiliar speech of the radio station playing in a taxi. Or in those moments of being overwhelmed and awestruck at the incomprehensible street in a foreign land. It can happen when we linger on the Chinese characters in a restaurant, rather than on their English translation. All that is required is releasing oneself from the pressure to achieve literal understanding, and letting oneself embrace a much deeper understanding.

He gets carried away with his own rhetoric to the point of verging on woo, but it’s still interesting stuff to think about, and that Dzongkha is indeed a pleasure to contemplate. Thanks, Jack!

06 Jun 05:55

Saturday Morning Breakfast Cereal - Consistent

by tech@thehiveworks.com
Yuval Pinter

"Testing Limits"



Click here to go see the bonus panel!

Hovertext:
Correlation doesn't imply causation! Induction is impossible! Hand me my briefs!


Today's News:

We now take a break from comics about aliens finding humans being weird to one about humans being weird.


Last time I wrote a post here I was surprised how many weirdos wrote in to tell me they still read the blog. Well, I'll just let you weirdos know that the amazing kids' book I've been working on for a year is finally (almost) heading toward the presses. I'm hopeful we'll be able to reveal more soon.

<3

Zach

01 Jun 06:11

קו האימות

by יובל פינטר

זוכרים שהיתה מגפת קורונה? דגש קל, כאן בשביל לעורר מחדש את תחושת הכיף.

לאורך המגפה ליווה אותנו הביטוי המיופמז-משהו ״חולה מאומת.ת״, שהתקצר ברבות הימים לפשוט ״מאומת.ת״. לכאורה, לפי התפיסה האפיסטמית שלי לפחות, מה שקורה זה דבר כזה: בת אנוש חושדת מסיבה כלשהי שהיא נשאית של נגיף הקוביד-י״ט, ואם נמצאה חיובית בבדיקה מהזן החביב עליה מצבה משתנה והחשד אומת. מעתה ואילך היא חולה/נשאית מאומתת.

אלא שאתםן קוראותים כרגע פוסט בבלוג הבלשנות המוביל בישראל דגש קל, ולכן נהיר לכןם שמשהו חרג מהתפיסה האפיסטמית שלי. ואמנם, שופופו את הפסקה להלן בהכתבה מתוך מעריב מספטמבר 2020 (לא, לא כאן נתקלתי בזה במקור, אלא ״איפשהו״ ומזמן כבר שכחתי איפה):

ואם ביום שישי נודע לי?

אלדד שביט מצוטט כאומר ״ביום חמישי לא ידעתי שאני מאומת״. בעיניי משפט כזה לא ייתכן, כי אם ביום חמישי הוא לא ידע אזי ביום חמישי הוא לא היה מאומת (יש כאן אולי פתח להתפלספות במקרה שבחמישי הבדיקה כבר התבצעה ואולי איזשהו גורם בשרשרת הטיפול בדגימה המזוהמת כבר ידע את העניין וסתם לא הודיע לחולה עצמו עדיין, אבל נניח לה כי גם מאנשים שטרם נבדקו בתאריך המוזכר כבר שמעתי את זה). מה שקרה כאן די ברור וכבר נתקלנו בו רבות הסמנטיקה הבסיסית של השורש א-מ-ת נגזלה מהמילה ״מאומת.ת״, שקיבלה חיים משלה כפחות-או-יותר מילה נרדפת ל״נשא.ית״. לא סטטוס הידע שלנו אודות הנגיף הוא החשוב כאן, אלא עצם הימצאותו בגופנו.

ונשאיר את קוראינו עם הציטוט הלא-בלתי-קשור-לחלוטין של מוסא אלפרון משלהי שנות ה-90: ״כבר 12 שנה שאין לי עבר פלילי״.

13 Apr 04:48

Saturday Morning Breakfast Cereal - Fungus Amungus

by tech@thehiveworks.com


Click here to go see the bonus panel!

Hovertext:
A tip of the hat to Patreon subscriber Brad Porter whose 'C is for Cordyceps' I cannot excel in this hovertext.


Today's News:
11 Apr 04:59

Please don’t send .docx or .xlsx files

by Kyle Gorman

.docx and .xlsx can only be read on a small subset of devices and only after purchasing a license. It is frankly a bit rude to expect everyone to have such licenses in 2022 given the proliferation of superior, and free, alternatives. If the document is static, read-only content, convert it to a PDF. If it’s something you want me to edit or comment on, or which will be changing with time, send me the document via Microsoft 365 or the equivalent Google offerings. Or a Git repo. Sorry to be grumpy but everyone should know this by now. If you’re still emailing these around, please stop.

01 Apr 09:45

Zero Days

by Doug
28 Feb 18:49

Saturday Morning Breakfast Cereal - Weakness

by tech@thehiveworks.com


Click here to go see the bonus panel!

Hovertext:
So ended the saga of Siegprinter.


Today's News:
21 Feb 05:13

Chomsky’s Forever War.

by languagehat

Geoff Pullum has a typically invigorating review of the new second edition of Randy Allen Harris’s The Linguistics Wars: Chomsky, Lakoff, and the Battle over Deep Structure; here are a few excerpts:

It is quite difficult to explain in nontechnical terms what triggered the linguistics wars, but let me try. Chomsky in the mid 1960s maintained that the structures of sentences were nowhere near as simple as the sentence diagrams of yesteryear; they were more like se­quences or layers of such diagrams. The most concrete layer, called the surface structure, captured facts about the overt shape of the sentence (word order, inflec­tion, and pronunciation). More abstract layers accounted for things like con­nections between sentences — relations be­tween active and passive clauses, for example. […]

However, from about 1967 some of Chomsky’s earliest defenders and most talented students began to develop arguments undercutting his case for “Deep Structure.” They claimed it had no theoretical necessity or significance. Instead, they posited much more abstract syntactic layers, and suggested that the most abstract layer of all looked much like what philosophers would call the “logical form” of a sentence. Their hypothesis was dubbed “Generative Semantics” (GS). […] GS emerged directly out of Chomsky’s work, using the kind of arguments he used, but it diverged from his own views, and displeased him. In his view this meant that the promoters of GS had to be defeated and punished.

The punishments Chomsky imposes on publicly announced enemies of his views are of two types: the dungeon and the fire.

Consigning an opponent to the dungeon of ostracism and obscurity is easier than you’d think for someone of Chomsky’s extraordinary prestige. His close followers, who dominate the mainstream of syntactic theory in America, tend to cite only what he cites, so if he resolutely avoids mention of someone’s name, within the mainstream it will soon look as if neither that person nor their publications even exist. (Today, when citation counts affect hiring and pro­motion, that can hurt more than the victim’s pride.)

The citational oubliette worked well for lesser-known or younger GS proponents, especially from low-prestige or foreign institutions. For example, in half a century the extensive and highly relevant works of brilliant European theorists such as Pieter Seuren and Esa Itkonen were never referred to in Chomsky’s vo­luminous writings, be­cause he had identified them early on as enemies.

But more visible proponents, especially those close to Chomsky’s home institution, MIT, faced something more like burning at the stake. Chomsky would immolate their ideas in classes (scores of visitors would travel long distances to absorb his wisdom at MIT each Thursday afternoon), in conference presentations and invited public lectures around the world, and in the many articles and books he produced. He excoriated, misrepresented, and ridiculed his enemies, sometimes in their presence, calling their positions incoherent, irrelevant, or possibly just muddled equivalents (“notational variants”) of his own. He would attribute to his opponents views they rejected, even when they were present and tried to demur. […]

By around 1976, GS proponents could see that they had lost. GS had been crushed in an ideological struggle be­tween an intellectual superstar whose followers saw him as infallible and a foolhardy clique that had dared to depart from the true faith.

Some on the losing side, worn down by a decade of struggle, looked like has-beens, which meant they could be transferred to the other punishment mode: total denial of citation. George Lakoff was tough enough to take Chomsky’s attacks, shake them off, move to Berkeley, and start a new life writing about metaphor and politics instead. But Chomsky’s former star student John Robert Ross was cowed and sidelined, and largely ceased publishing on syntax (a pretext was found for forcing him out of his tenured faculty position at MIT). James McCawley continued writing GS-influenced works from his indepen­dent base at the University of Chicago but was increasingly ignored by the main­stream. […]

Harris quotes Shakespeare’s amusing vignette of Hamlet persuading the sycophantic Polonius that a certain cloud is shaped like a camel, then getting him to agree it is like a weasel instead, and then that it resembles a whale. True Chomsky followers show their Polonius-like loyalty through unquestioning acceptance of each new image Chomsky sees in the theoretical cloud. They adopt the buzzwords, modify the direction of their own work, and pursue the new line — only to be left high and dry eight to ten years later when once again he demolishes the central pillars of the previous conceptual edifice.

Another practice of Chomsky’s is to plunder the work of his opponents for attractive ideas and useful terms to be adopted without credit a decade or two after he first attacked them as misguided. In recent years he has mined the program of GS for insights without crediting the original developers. Around 1990 he completely abandoned his “Deep Structure” level without a word about this being the central proposal of GS in 1967. His “minimalist program” posits a unified rule system relating logical forms directly to surface forms of sentences, a position GS proponent Paul Postal explicitly advocated in 1969, when Chomsky hotly rejected it. Chomsky never mentions that adopting the phi­loso­phers’ term “logical form” for the most abstract level of sentence representation was George Lakoff’s idea.

It’s hard to stop quoting, but if you want more just click on through; I know I’ve bashed Chomsky a lot and quoted a lot of bashing by others, but I was traumatized by him and his minions at a formative age and can’t get enough. I got the link via Mark Liberman’s Log post, where David Marjanović left this comment:

The latest shot in the linguistics wars: an open-access paper on how language can indeed be learned from input alone.

Venceremos!