Cultural AIppropriation

  

In 1997, there was an outcry when ad people made an a commercial in which Fred Astaire danced with a vacuum. It seems that in the real life filming event he wasn’t actually dancing with a vacuum, but with a cane (!) so people didn’t like the idea that a computer system could fabricate images. Now, in the era of deep fakes and the like, the problem has, as you might expect, disappeared. Aging and de-aging actors so that they can play other stages of life, and using CGI to insert actors who died have become acceptable ways of generating relevant cultural content. I also keep seeing ads on my various feeds for a service that can create voice-over avatars, inventing the semblance of a pretty face who can lip-sync my words.

And then I saw this https://typecast.ai/learn/ai-voice-cloner/#:~:text=AI%20can%20clone%20and%20modify,custom%20voice%20perfect%20for%20projects.

This system can, apparently, take my voice, clone it and have it say things I never said. The ethical questions are staggering:

https://www.forbes.com/sites/lanceeliot/2022/07/02/ai-ethics-starkly-questioning-human-voice-cloning-such-as-those-of-your-deceased-relatives-intended-for-use-in-ai-autonomous-systems/?sh=4110e3ad4882

While I know that there is much to argue regarding the morality of manipulating people by impersonating their loved ones, but I am going to stick to an application of a recent issue – that of appropriation.

I am a fan of the Simpsons and I have been since before the first episode (I was a fan of Tracy Ullman). One character in the show is named Apu and he was voiced, for a long time, by Hank Azaria. In 2020, Mr. Azaria decided to stop voicing the character because his imitation of accent was considered offensive. For you Family Guy fans, feel free to retell this story just using Mike Henry’s voicing Cleveland Brown until he didn’t anymore. Now, on one hand, I can rail against this because actors’ impersonating characters from other places is pretty much standard. Heck, on the Simpsons, Bart’s voice is rendered by Nancy Cartwright and Erin Chase has voiced Charlie Brown (and Dan Castallaneta who voices Krusty is not Jewish). Watching the spate of British actors (Hugh Laurie, Benedict Cumberbatch and Tom Holland, I’m looking at you, mostly because you are all really good looking) play Americans or Jared Leto play an Israeli, Oscar Isaac don an imperfect English accent, Toby Jones pretend he is Swiss or Sacha Baron Cohen play most any role is part of the suspension of disbelief. But on the other hand, an outcry is an outcry, right?

Now, this isn’t really all that new. Remember that Freeman Gosden and Charles Correll, 2 white men, played the voices of black characters Amos and Andy. But times are apparently different now, as they often are.

So, first, let’s consider where we draw lines. Blackface is out. Jewface is in. Womanface is in, but only sometimes. Appearances are still in flux. Authenticity (that is, the actor’s matching the character’s particular characteristic) is highly variable. A straight man playing a gay character and vice versa? I think that’s still acceptable but I could be wrong. A Hispanic actor playing a non-Hispanic historical personage is OK as long as the entire story is a treatise on race and ignores the actual reality behind the story. ScarJo cannot play a Japanese character as non-Japanese. Anne Boleyn was not black but casting Jodie Turner-Smith is ok. And at this point, I’m not sure if Nick Fury is black or white.

But be all that as it may, and it may, my question today is about the use of AI to play roles or present voices that are inauthentic for about a million reasons. Can a voice-cloning AI use a celebrity voice to play a role? What about an effected accent or a comedic take on a voice? Could a programmer use the past performances of Mr. Azaria to create a new performance of Apu for future episodes? Who would get the credit and who the blame?

If a computer cloned a pidgin or a dialect which is less than/different from the king or queen’s English, who would be committing the offense? The programmer? The person who spoke in the sample that the computer ate in preparation for spitting out new words with the old voice? Beyond the issue of intellectual property laws which would come into play if a company uses a voice without permission, creating the impression that a famous person endorses what he doesn’t (this has already happened with celebrity impersonators) or the voice of a dearly departed celeb (without permission of the estate) there are the questions of that confounded authenticity. The AI isn’t Jewish, Hispanic, Asian, Brooklyn-ish or whatever. Would an AI vocal performance employing the stereotypical affectations which make our pap culture palatable be criticized for relying on the same shorthand that our actors employ? Can computers be attacked for appropriating the culture of others, any others – because a computer has no native culture? What if the programmer is of that culture?

We need to hate someone and the AI isn’t a someone and this might prove a problem when we try to channel our righteous indignation.

 edit -- https://www.nationalreview.com/2023/06/let-the-beatles-be/ interesting comments on AI's taking over a voice

Book 'em

So let's just say this right out front -- I'm not in favor of book banning. But, truth is, I don't think most people are. I believe that there are a number of other approaches to books that require analysis but aren't as absolute as forbidding a book's being published, or collecting books that exist and eliminating their presence from the greater universe.

I think that if you were to get together a large group of people who are "liberal" minded, whatever that means, and ask them if it is appropriate to hand an 8 year old "Tropic of Cancer" most would agree that the book is not intended for, nor healthy for a child of that age. We can agree that some things are appropriate for people of a certain age. For the most part, people won't object to the "voluntary" rating of movies. (I say "voluntary" because IIRC an unrated movie will have a harder time with its mass-market distribution so most movies made for the public-at-large really HAVE to submit to the MPAA to be commercially viable).

So if we can agree in principle that some things are ok, or are not ok for people of a certain age, then the rest is just a matter of working out details. Trust me -- this sounds horrible and, as a younger man-child I was appalled at the efforts of the PMRC to "rate" music. But as a parent, it has become clear that I would rather my child not hear certain words and concepts masquerading as music so some sort of advisory on the outside, letting me know what the child is getting in to is a good idea. Additionally, content warnings on TV shows (and the ability, theoretically to adjust a V chip or equivalent so that youngsters can't watch stuff), and even age confirmations on web sites, or filters all are good things.

Now, maybe I'm not understanding the attempts made by certain forces so I'm using language differently, but I think that an equivalent rating system on books is desirable. We have quantifiable standards that allow experts to decide what music, tv shows and movies are appropriate for what age. Why not have the same standards established (violence, adult themes, language etc) for print texts? And, yes, then I think that a school library might choose either not to hold a book which would not be something for children to read, or to have the books that are for certain ages be kept in different locations in the library apart from those for more tender eyes.

But shouldn't the parents have the right to decide what their children read? Sure, just as an adult can walk an 8 year old in to an R rated movie, an adult can buy a book, or take it out of the local library. But that doesn't mean that an R rated movie has to be shown in the school. And then the question becomes "why do my taxes support a public library that stocks books X, Y and Z?" and to that, the answer is that we don't have line-itemization in taxes and my money goes to all sorts of local functions and programs that I might not otherwise support. Our libraries don't stock every book -- some things are, indeed, too (insert controversial content here) and librarians and their communities have been making judgments about what to have and not have for a while now.

So, again, I'm against anything akin to a complete ban, but I do think that all stakeholders -- parents, teachers and even politicos can help determine what is properly accessible to whom and what isn't so that in certain contexts. Will we all agree in each case? Can we set certain universals (ideological issues aren't subject to ratings -- we limit to descriptions of violence, sexuality and scatalogical imagery maybe?) as a working-starting point? Will it be easy? Will it be quick? Will it be free?

But does a challenge mean we don't try if there is an underlying point of agreement?

Aging out

 

I never thought I would be the guy who wrote about defending standardized tests, but here we are. I also never thought I would appreciate Taylor Swift, but, like an alligator, she has her time and place.

My usual position is that standardized tests

a) have test maker bias built in

2) can't assess for process

iii) can't assess for skills (or assess only for a narrow range of demonstrated orders of skills including "test taking)

four) assume all students develop identically

⬠) allow grading bias (for free responses)

0110) reward an economic class that can afford in time and money effective tutoring

and I'm sure some other problems. But here's the thing -- I also recognize their utility. As I explain to my classes, X University has no way of comparing a public school student from Duluth and a private school kid from Paramus unless both sit for an identical assessment. Our curricula differ. The socio-cultural values and influences differ. And on and on.

So, flawed as they are, standardized tests, if used properly and considered properly are useful. OK, file that away.

I also believe that a person's age is important in deciding maturity and readiness for a variety of things. Is setting a particular age-limit on anything just begging for exceptions? Yes, but on the whole, it is clear that a 13 year old isn't ready to drive, or get married, or judge candidates maturely enough to vote and a 23 year old is. So we dicker over specifics but agree in principle. I think that the same should hold true at the other end of the spectrum -- we already do this by having, in some industries, mandatory retirement age. This not only allows an organization to phase out higher-earning veterans and cycle in new, cheaper hires, but, theoretically (and a bit less cynically) it establishes the notion that competence wanes over time. And for certain skill areas, I think we all agree that this is true. 

The fact that it is biblically precedented need not be mentioned here, so I won't.

So it might be reasonable to say that, because of the majority of cases, jobs that require heavy lifting are not for people over whatever age the experts in the field see a person's needed strength ebbing. And I don't make that judgment. The age parameters need to be established by experts in the field because I don't want someone whose peak performance is limited by the frailty of age performing a task that could mean the difference between life and death for me. Or some people I know, even.

Therefore, I am proposing age-based testing for a whole lot of stuff. Let's start easy -- driver's licensing. In the same way that you need to take a written and practical test to earn a license, starting at a certain age, testing to maintain licenses should be required. Vision, hearing and response time. 

Is there concern over the thresholds that can be established for competent performance (and aren't they rendered moot by the advent of self driving cars)? I think that baselines can be set up by conference of medical and performative experts. We have a limit called "legally blind" so we know what you need to be able to see to drive. We can determine what one needs to be able to hear when driving, and set that as a limit (if there is no current standard requirement for hearing when driving, then scrap this branch). I think we can also create a standard for response time (I know it is tested to show people who are under the influence the difference between their response time and a sober driver's). 

Is there a concern over cost and implementation? There need not be, not because we already have an infrastructure built, that for road-tests for new drivers, so we could just put people of an age or higher through the same process, but because we also have new technologies. Simulators can be used effectively here because the goal isn't to assess the fundamental skills of driving, i.e. the rules of the road, but just reaction time and critical judgment skills. I assume that this will knock certain people out of contention for holding a license which will drive down insurance rates for everyone, take car-impediments off the road and increase work for drivers of the Uber sort.

Next up, gun licenses. I think that beyond a certain age, when the same (2 or) 3 components degrade, one should risk forfeiture of licensure. And I think that vision, hearing and cognition should be tested to ensure the safety of everyone else. Heck, we should be testing for mental competence from the get go but I don't know how that plays in.

Do we want a policeman, or a fireman, or an EMT who cannot carry, run, see, hold steady or judge a situation properly anymore? We work from the list of desired outcomes we want to check on and work backwards, designing practical (and theoretical) scenarios to ensure continued ability. And, yes, I think the same holds true for politicians. While vision might not be the same sine qua non, the health-physical should be complemented by a mental-acuity assessment yearly to make sure that memory and discernment are still up to snuff. Supreme court justices, at a certain age, aren't thinking as clearly as they once did (though their positions are as much the function of blind political ideology or the efforts of their clerks as they are of the justices' own input so they can stay in their jobs longer because, ironically, less is demanded of them).

Look, I'm aging (EVERY DAY!) and I know that at some point I will be nothing more than an adorable teddy-bear in the classroom, not able to think as quickly as my students, not able to command their attention or handle the physical, emotional and mental rigors of my job. I also know that while the risk to students is limited and that the real crime is defrauding parents of their tuition dollars by sticking their kids into a room with someone unable to provide any return on investment, still, my usefulness in the classroom will eventually be in question. So is my job "safer" for an older person than a job in the field of "bomb disarming technician"? I would assume so. But I'm not looking for a sinecure in which I can do least harm because parents should be looking to fund that. Will I hold on to certain skills even past my freshness date in the classroom? I assume so. The idea that those who can't do, teach, is not wrong. Someone who is past his prime physically can still transmit experiences and lessons learned through mentoring, so there are still spots for those who are not field-strong anymore.

So, yeah, I think we need a combination of age-gates and standardized assessments. I know it will cost money to create and implement and that the results will (as with all assessments) be inexact. But we will also be working to make sure that the productive engine of society continues to be strong.

More, LOTS more

  

Back when I was but a boy, still optimistic about my ability to learn math, my teachers taught us the concept of “estimating” (also known as “guestimating”). The idea was that we could approximate what the answer should be like before we did any computations so we could compare what we got with what made sense and that way, we could see if our math was completely off base. At the time I felt like the estimate should have been enough because, even then, I knew that there was precious little use for math in the world. Leave that up to the professional mathetizers and I’ll stick with the lottery winners and hire them mathetizers to do the work.

As I aged (for lack of available and desirable options) I recognized that the value of estimating is that it reinforces an understanding of the relationship between numbers. Once you see how the answer SHOULD look because you have a sense of how the pieces operate, you stand a much better chance of seeing why a potential answer is wrong. If you know the answer should be a negative, or has to be more than either of the parts, then an answer that doesn’t fit that schema can be discarded.

This type of understanding of numbers is rudimentary, but, more importantly, it is fundamental. If I go to the local fast fooderie and pay with cash, I don’t want my server to take my cash and give me change without considering what, approximately, my change should be. I want to hear him or her say “that can’t be right!” Otherwise, typing errors or other tech-interface mis-actions will be believed regardless.

Why do I mention this? I’m glad I asked. I’m also glad I didn’t wait for you sheep to ask because I can’t always count on you people to wade in and find your voice. So I’m here for you. Mostly because I’m awesome.

I was having a conversation with someone recently about the value of math instruction and he pointed out that he didn’t like when, when he was but a wee lad, his teachers forbade the use of a calculator during assessments because, in their words, “you won’t always have a calculator handy.” He laughed at them in retrospect and reminded the huddled masses that he always has his phone with him. Har har we all chortled (chortling is the cool way to laugh kids). And secretly I cried inside.

What I predict is that people in the service industry, doing the kinds of basic work that serves as the backbone of any society will start doing worse and worse jobs. The guy basing his fee on the square footage of my living room floor will use his phone to measure and then compute and if the resultant acreage (It’s a big room) is illogical, he will trust the technology. The guy figuring out how many miles I can go before my next oil fill will mistype a number into his device and I’ll end up on the side of the road next to a flaming hunk of imported metal.

The same can be said for teaching spelling – understanding WHY words are constructed the way they are not only helps sharpen the memory, but allows people to intuit the meaning of unfamiliar words by looking at roots and affixes. Relying on AI to “correct” words will lead to the wrong words’ being used (the “Damn you autocorrect” syndrome) and the wrong message’s being conveyed. We cannot dress up laziness and call it “the future” and turn it into a positive.

The Lower Order Thinking Skills which I value so highly, having been subcontracted to technology will disempower the sectors of society which are based on those skills and there will be no intuition – no developed memorization skills, no experience of spotting the error, and no understanding of the relationship between values. There will be no practice of approximating a proper answer because “who needs to do basic math?” There will be no insight into WHY math works the way it works and therefore there will be no critical sense willing to question when an answer doesn’t fit, on its face.

We will not be taken over by robots, technology and AI. We will be (and already are) handing over the keys to the kingdom in a slow regression that we call progress. We will become reliant on technology and so convinced that our deeper analytical skills are so developed that we needn’t worry about the basics. And the robots won’t have to do much but exploit our foundational stupidity and we will go to our doom in neat and orderly lines.

Confusing Food With Food

I want to create a menu. I do love to cook but this is a menu of foods I am not going to make, but that I would like others to make and then take pictures of. You might even want to eat it. Don’t know, don’t care.



Here is what is driving the menu: wrong color foods.



Ok, so the colors are not “wrong” per se, but are certainly unexpected.



So, for example, a plate of spinach fettuccine (or for non-kosher consumers, squid ink pasta). Got it?



Let’s get started:



At that table there should be a bowl of techina made from black garlic seeds and a brown butter/black garlic oil. Also, golden grape jam for spreading.



Start with a pasta bar with 3 choices:

Squid ink/black

Beet pasta (pink/red)

Spinach pasta (green)



To pour over it, we have a choice of a green cheese sauce or topped with a tomato sauce made from yellow tomatoes. Such things exist, or so I have been told, and I imaging the contrast of a bright yellow sauce with a green (or black) pasta would be striking. Or a green cheese sauce on a green pasta might just be gross looking!

Sushi made from Lingcod

I want a corn salad made with purple onions and red corn. Yes, red corn is a thing. Maybe chop up some non-green peppers. Not as unconventional, but tasty and colorful.

Wax beans on the side and a white pumpkin pie, plus some grilled purple cauliflower and purple broccoli, and white and purple asparagus.

The corn, alternatively, could be used to make a red/blue creamed corn or a red or blue corn-based corn bread, tortillas, or chips to go with the dips above.

Next, a tzimmes made with purple/blue and red carrots, or maybe just white ones. And a mashed potato dish made from purple potatoes, but still savory.

Slice up some white sweet potatoes very thin and fry them up.

Serve with a glass of blood orange juice.

For dessert we have a lot of options:



Dried rings of pink pineapple

Slices of white or yellow watermelon

Golden raspberry compote

Red kiwis

White cherries and white strawberries served next to a chocolate fountain filled with white chocolate

I was considering a red banana pudding but the pinkness of the red banana flesh seems fleeting.



So start shopping, cooking and taking pictures of the meal that will mess with your senses.

My official position on the use of AI in the classroom

 

Yes, here it is. I have reviewed my thoughts, catalogued my experiences and quantified the qualifiers, and here it all is.

First, the question: how do we, as instructors of writing, hold students responsible for writing literary responses when there is the specter of AI writing to contend with?

I think that this requires that we re-evaluate exactly what we are trying to accomplish in our instruction of writing and the writing process. I have dealt with many of these issues here and here and I will be repeating some earlier sentiments (apologies) but then crafting them into practical applications by the end of this piece.

As I have written elsewhere, I am a fan of establishing a strong foundation of concrete skills for students to rely on and build upon. We cannot absolve students of the need to be able to do simple arithmetic just because they have access to calculators. Not only will there be situations in which the calculator is inaccessible, but the thinking skills (read: neural passageways) that are developed train the mind to be able to do other types of thinking -- it is an exportable strength. The patience and strategizing-organization that the brain must employ are not limited to doing simple arithmetic, but to all problem solving situations. So the skills, important on their own, are essential in conditioning the brain, preparing it for unforeseen challenges. Stretching before a basketball game makes the muscles more capable on the run home as well.

I reiterate, then, that students still need to learn, acquire and master skills and in my 10th grade (grade level) class, this means mastering a particular structure for a thesis, a body paragraph and a transition. To me, these are non-negotiable elements to proper and complete student writing. Sure there are many ways to go about structuring essays and thesis statements, but in my class, I use a specific form and formula. The resultant essay is "cookie cutter" but it is complete and does the job. AI cannot teach these skills so my job is somewhat set. And if I'm concerned about the AI involvement, then I can (have and will continue to) require students to practice these skills in the absence of technology. Yes, this smacks of hiding my head in the sand, but I don't think that we can abandon the ol' pen and paper. Unless a student has fine motor skills issues, I believe that any students can write (or dictate, in a situation in which the accommodation is a scribe) a sentence that follows a formula. I believe that a student whose accommodation is the use of a computer can be monitored in the classroom as he types a single paragraph. Either a lock-down browser or simply an observant teacher and a well-placed computer can dissuade a student from using an automated system instead of his own head. Students (at least on this level) need to be drilled repeatedly, tested to see that they can make a thesis, build a paragraph which connects text and idea and then set up a transition which reaches back to complete a thought and introduces the one next expected via the thesis.

Earlier today, I sat down with the ChatGPT website and fed it one of my essay prompts, verbatim. I then had it generate 10 "different" responses. I put that word in quotes because it often churned out identical sentences and phrases (in fact, were I to try and detect whether an AI composed a response, I would first run the prompt through and then look for repeated phrases and structures in the responses I evaluate because the LLM repeats itself). What the computer came up with wasn't horrible but it could not predict structures that I taught in the classroom, even if I used phrases like "proper thesis" or "transition." The writing was generically competent and even (mostly) textually accurate but would not satisfy the requirements of my class. Paragraphs lacked specific detail, included over-generalizations and external information and relied on broad statements. The ideas were left unconnected. I believe that a student, trained in how to create a thesis, a body paragraph and a transition in my class would see clearly the failings of the AI generated response, and if not immediately, would see them after more direct instruction to help him or her acquire this specific skill.

The solution then (in my not-so-humble opinion) is to isolate the two categories of skills and address each one differently: students should be required to construct the necessary elements of a literary essay without technological mediation. Simple as that -- if we are concerned about the use of AI, require that all work is done in class and without computers (or without unmoderated computer use). This will address issues of Lower Order Thinking Skills and will ground the student in the relevant and class-established expectations.

AI provides the opportunity, though, for the evaluation (a Higher Order Thinking Skill) of text and frees up the student from HAVING to create the material to be evaluated. This is its power -- it speeds up a process that we could do if we wanted. If we can't do it without a computer then we shouldn't be relying on technology. But if we can, and just want to move on to other things, then the AI is a great tool. The real problem I'm seeing is that people use technology instead of developing the skills so they can't evaluate the product created outside of their brains. If I am used to having a computer compute my averages in class but I don't udnerstand how those averages are computed than I won't see the problem when the computer generates a result which is wrong. If I can't read a map or plan my own trip, then I can't assess whether computerized directions might be in error. Attaching motors to pump my legs if I don't know how to walk won't make me a faster runner -- it will cause me to fall on my face.

The real skill that we have to cultivate then, after we get through the basics of content creation is content revision, and by using the AI to make the material to be revised, we can craft tens and hundreds of equivalent samples that students can then work on. The student can then practice the higher order thinking skill of evaluating and revising, assessing and commenting without having to spend the time creating the base materials with which to work. But this only works if the student understands how to make that basic content so he has something to compare the AI creation to. In the actual work by the student, the technology is unnecessary as the practical skill of revising is easily accomplished on paper so we no longer have to worry about the reliance on AI. And if the assignment is on revision, to ensure that it is done authentically, we can collect the work, not just the finished, revised work. We can say, like a math class, that the student can use a calculator but we have to see all the work.

The assignment, then, has a variety of steps and two distinct components:

1. Assign a text

2. Assess the reading (through quizzes, short responses, class discussions)

A. instruct on the format of a thesis, body paragraph and transition

B. practice elements of item A repeatedly

3. Assign a lockdown-browser, or by-hand timed assessment requiring elements in item A

THEN

4. Feed the prompt(s) for step 3 into an AI text generator

5. Have students "grade" the product of step 4, comparing it to the obligations expected because of item A

While larger writing assignments can then follow, especially if they are based on work generated in-class, or even AI material revised by the student and brought up to class-spec. If we are starting a larger assignment from absolute scratch, and we have no particular format or stylistic demands, then we might require running the completed product through an AI checker (which I don't particularly trust), but at that point we are as concerned about the student's honesty as we would be of a student whose tutor does most of the writing when essays are assigned to be completed at home (which cannot be caught by a plagiarism checker).

In any case, we free ourselves of the ChatGPT of Damocles because we have established that the student has the actual skills expected and then can apply them in an authentic way.

Teaching writing is about teaching the process of creation so we need to evaluate the steps in the process and stop being worried that a final result which materializes out of thin air is suddenly suspect. There is no assessible process when the assignment is left to be done away from school, begining to end so that is a less effective measure of any specific process-based skill anyway.

True strength

 I was driving in my car last week, headed to the airport to pick up a child and a spouse. Newark airport has redesigned terminal A so I'm still getting used to the new layout. Here's a weird detail -- as you approach the terminal there is an entrance onto the road from the left side and it isn't one of those long ramps that allows cars to merge; it is a 90 degree left turn into the fast lane. I was unaware of its existence until a large car decided to inform me by lurching all the way across my lane and half of the next lane over before beginning his left turn. I swerved to the right, hoping that, early in the morning, there would be precious few cars alongside me. I also hit the horn (which I rarely do) and slammed on my brakes. This adventure in multi-tasking had a number of effects. First, I (AFAIK) did not die, so there's that. But I also found my right hand splayed across the passenger seat. I had no conscious thought of doing this, but the instinct to protect a passenger is very strong in this one. Had there been a porcupine in that seat, he or she would have been safe but I would have become a pin cushion.

A similar instinct kicked in when I walked my kids outside. I would hold a daughter's hand simply because, heck, I happen to like my kids. But if she were to stumble at all, I somehow (possibly magically) instantly squeeze her hand to keep her from falling. No one taught me to do this. I don't think about it, I just do it.

So when I was reading the haftarah from yesterday, and I encountered the story of Uzzah, I became, shall I say, "perturbed." Yes. Yes, I shall.

Perturbed.

In the story (2 Samuel 6:6), the ark of the covenant (cue Indiana Jones music) was being transported via wagon to a new resting place. Uzzah, son of Avinadav, walked along side it, and when, at one moment, an ox moved in a way that caused it to shift, Uzzah instinctively reached out his hand to stop the ark from falling. But, by touching what he was not allowed to touch, he incurred the death penalty. It doesn't seem fair. All he was trying to do was protect God's honor - you can't let something so holy fall on the ground. And on top of that, he probably didn't stop and think; he just acted. Why should he get punished for something out of his control, and something driven by the holiest and purest of desires, to safeguard the ark?

I did a little digging and found that I wasn't the only person troubled by this. And each person, sometimes citing some sage or another, sometimes speaking on his own, tried to explain what it really going on. Here is one example https://forums.dansdeals.com/index.php?topic=60155.20. And I get it. After a bunch of reading and thinking, I can derive a nice homiletic explanation which applies to our daily lives et cetera, et cetera. But I still don't like it.

First, a piece of background (and I'm getting foundational info from the Artscroll Chumash, page 1168. There are more formal citations on other webpages so feel free to check them out, but I'm not waiting for you).  The ark was not supposed to be transported via wagon. The commandment was for it to be carried. In fact, the medrash learns out that the ark carries itself and sustains those who carry it, so they do not need to worry that it is too difficult. It does the heavy lifting. Putting it in a wagon might seem more efficient, or simpler, but it isn't what was commanded. Uzzah let that happen. He didn't defend God's honor by championing the actual obligation of carrying. He stepped up to "protect" only after he had shown that he wanted to be next to God but only on his own terms.

So in this moment, Uzzah forgot that the ark didn't "need" him and he felt that he, as a human, was essential in keeping God "safe" (as it were). That's a message that I have seen on websites. But I think it goes beyond that. I think that the underlying message is about the need for awareness and intentionality, even when it seems that instinct is just as effective.

When I pray, am I aware of the words that I say? Each moment? Every time? Or do I say them because "that's what we are up to" and I have said them so often that it happens automatically? When the chazzan says kaddish, do I respond "yehei sh'mei..." because I am thinking, at that second, about actively and consciously praising God's name, or because I click into automatic and respond? This doesn't mean that I don't mean the words as I say them. It means my decision to do it has stopped being a decision. And when it comes to protecting my little baby as she takes her steps, or my passenger as I drive to the airport, that automatic response is perfectly appropriate. But in our approach to the divine it isn't enough. Uzzah wasn't wrong in his motivation, but he also wasn't acting from within an intentional schema. He lunged because he saw God's ark just like anything of value and we act without thinking to protect anything valuable. Admirable but not what God wants. God doesn't even want us to do things pro forma -- with conscious intention to do them but without the thoughts surrounding our actions as a proper decision to express a specific and unique emotional connection at that moment. Think of Shmu'el statement to Sha'ul in 1 Sam 15:22 -- does God desire sacrifices? He wants you to listen to what he actually commands. Don't do it because YOU think that that is what is necessary. We should do what GOD says is necessary.

Uzzah, consciously or not, was doing what HE thought was right. Now, maybe it is hard to blame him. It is tough to defeat instinct. It requires a level of personal awareness which is really (really) hard to achieve. Some huge percentage of our actions are either purely instinct (try NOT flinching when the ball comes straight back behind home plate, even though you know there is a net to protect you) or learned responses which eventually border on instinct. And there is a reason that the major league catcher is such a prized commodity -- he can control his responses so that he doesn't flinch in that crucial moment. Uzzah was acting like a guy in the stands. That's great if all you aspire to is being a guy in the stands. But if you want to be a major leaguer, you have to be aware in that moment to such a degree that nothing you do is instinct even though it seems to be.

When we pray, when we put on tefillin, or when we give tzedaka or do any other mitzvah, are we doing it

a) automatically because that's what we do

b) as a learned response because we figure that's what we are supposed to do

c) because in that moment, we have slowed time down, considered all the options, and realized that this is what we are commanded to do

Uzzah acted according to A or even B. He needed to act under scenario C. No, it isn't easy. No, it isn't fair. But if we want to merit walking carryng the ark, then we have to be ready to shoulder that burden and take every step consciously.