Memory is more than Ebbinghaus

This blog is inspired by the #blimage blog series.

“Dante Alighieri01″. Licensed under Public Domain via Wikimedia Commons

In February my wife and I visited Florence in Italy. It’s an extraordinary place – go for the art, go for the leather shops, go for the food, but if you ever have the opportunity, just go.

Walking past the Uffizi we paused in front of a statue. Recognising it was Dante Alighieri, we fell into a collective fit of the giggles. Then we recited together:

Dante Alighieri
Seldom troubled a dairy.
He wrote the Inferno
On a bottle of Pernod.

It’s a Clerihew, named for its inventor Edward Clerihew Bentley. All Clerihews follow the same four-line format and contain some ridiculous biographical nonsense.

My father – who is fond of all forms of poetry – recited this Clerihew to me when I was a young teenager:

Sir Humphrey Davy
Abominated gravy.
He lived in the odium
Of having discovered sodium.

I heard this Clerihew once, and remembered it. In contrast I had to ask my dad several times who the author was before I could remember his name – Clerihew. The picture of Davy suffering at the hands of the mob for his science was at once ridiculous and vivid and formed a dramatic picture in my mind.

Years later, living in Turkey, I purchased The Complete Clerihews at a British Council library sale of excess stock. A quarter century later, it remains a treasured book, always guaranteed to produce a laugh with its combination of Clerihew’s text and illustrations.

It is from this book that I read Dante’s Clerihew. Again, it stuck at one reading – it was so deliciously absurd, how could it not? At some point in our near two decades of married life, I recited it to my wife who laughed, appreciating in particular the sly innuendo of ‘Never troubled a dairy’ and it stuck with her, too.

And so, a few years later, we found ourselves in Florence, laughing at poor Dante’s statue.

All this says something about memory. Two different people hear 15 words in a particular order and remember them perfectly at one sitting.


Unlike my memory of Sir Humphrey Davy’s, I have no strong mental image of the Dante Clerihew, so this isn’t about pictures. Its appeal is purely in the sheer delight in the words, the absurdity of the situation, the portrayal of Italy’s leading man of letters as a lush, the anachronism of the drink Pernod. Together those 15 words form a coherent whole that stuck a vibrant chord with my mind – and my wife’s – and stuck, seemingly forever.

Now contrast this with what is often repeated about what we know about the human memory from the experiments of Herman Ebbinghaus. The ‘Forgetting Curve’ has been a standard inclusion in L&D presentations for many years now (including my own):


The Forgetting Curve

The common interpretation of this curve is that memory automatically decays after we are exposed to something new, and it needs several repetitions to remember the new thing.

It’s actually more complex than that.

Remember how Ebbinghaus did his experiments – by reading new, nonsense syllables and attempting to remember them. All subsequent experiments to explore this phenomenon have taken the same approach. The syllable sets are carefully sifted before they are given to people to ensure that each individual non-word has no meaning.

This is entirely right, if you want to explore people’s memory properly, you can’t have some people better able to remember their syllables because the syllable (let’s say ‘BRIK’) sounds more familiar than another group’s syllable (say ‘ZIK’).

In memory experiments, the content you learn is meaningless.

But in real life, much of what we learn is not meaningless. When my father rattled off the Humphrey Davy Clerihew, I already knew who that famous scientist was; I knew he had discovered sodium, and I knew what odium meant. There was very little entirely new meaning that I had to take on board. Only the arrangement of the words was new. The Forgetting Curve did not apply.

Let me be clear here. I am not saying that the Forgetting Curve is wrong. Quite clearly it is right for a great deal of what we learn. In fact, the very word Clerihew required me to do some spaced learning (going back and asking my dad repeatedly what that bloke’s name was) until it stuck. Why? Because the word by itself had absolutely no meaning. There were no hooks in my mind ready for it to latch on to.

In contrast, when I read the Dante poem I knew all the words, and they made me laugh. If I had been asked to remember this, instead:

Dairy the Alighieri
Pernod troubled on he
a bottle Inferno
On wrote a Dante bottle of seldom

I would have once again required spaced learning – repeatedly testing myself until I got it right.

Not everyone is a fan of Clerihews. But everyone has a memory that is particularly attuned to learning some things very easily. For me, it’s song lyrics, the origins of words and useless facts like the speed of light in miles per second and the capital of Mongolia.

In addition to these idiosyncratic traits, people develop the ability to memorize certain things faster and more accurately as a result of their work. Famously, Chase and Simon (1973) showed board positions from chess games to a range of players from beginners to international masters for about 5 seconds. When asked to recall the location of the chess pieces, the beginners could manage on average 4 pieces, while the masters could recall virtually all 20 pieces.

This was not a matter of super memory but of meaning. When shown a board not from a game, but with randomly arranged pieces, the masters fared no better than the beginners. They were unable to use their immediate appreciation of how the pieces related to each other to make sense of the position. It was like trying to memorize a poem with jumbled up words.

This knowledge can be used in learning and is. Sometimes spaced learning is the right choice. Sometimes it is the only choice. Sometimes we make content vivid, we use humour, contrast and absurdity to make it need less repetition. But as well as these tricks of the trade, the best content designers do something else.

They go to where the person already is, and use what they already have, to make the content coherent, resonant and easy to remember. They present their chess players with a real game position; they give their poetry lovers words with rich meaning, because they know the best language – the only effective language – is the language of your audience.


If you’d like to know more about spaced learning, read Dr Will Thalheimer’s excellent research piece.

If you’d like to take up the #blimage challenge, pick one of these photos and see what it triggers in your mind:

2015-05-12 15.15.56


2015-04-30 16.47.40

2015-06-03 13.52.51

Do you want to make a difference in L&D, or do you want to be liked?

June is an important month for Learning and Development (L&D). It’s Debunk Learning Styles month. It’s important for the profession, and it’s important to each of us individually in L&D because it brings up a key question.

Do you want to be liked, or do you want to make a difference?

Continue reading

The limits of Twitter – trite, glib and lazy

We can expect lots of Twitter activity in the run up the UK election on May 7th. Here’s an early shot in the skirmishes:

Shocking, right? The problem is, it’s not true. Not in the literal sense of the words.

30% of nominees to the House of Lords contribute 97.9% of donations to political parties.

Literally, that means that the vast majority of money given to political parties comes from 30% of nominees to the House of Lords. A moment’s reflection told me this couldn’t be the case, so I went back to the source, a 33-page paper from Oxford University: Is there a Market for Peerages?. I also scanned reports on this paper in the press. Here’s how the Guardian put it:

That left 92 “others”, who donated between them 97.9% (£33.83m) of all the donations coming from nominees to the Lords

That’s clearer. It’s still pretty shocking. As the Daily Telegraph summarizes: 92 nominees to the Lords had effectively bought their way in. (Got the cash? Interested? Best bet is the Tories, accounting for almost half the 92 nominees.) However, the point is also that this is not all money donated to political parties. The 97.9% only refers to the share of money donated from among those nominated to the Lords.

The problem for Twitter is this: the Guardian’s phrasing may be clearer, but it’s 5 words too long, and a little too complex. What to do? Cut. Never mind the accuracy, feel the brevity.

This is why I’ve been using Twitter less recently than I did in the past – it’s just a little too noisy and a little too inaccurate.

To be precise: there is always noise, but the noise to signal ratio seems to have increased significantly over the past year. Also, among the noise is too much of the glib and the trite. The glib includes motivational quotes that irritate more than they motivate. The trite manipulates the truth to lazily present an inaccurate picture of it, which is what @Election_HUB seems to have done with this fact about buying peerages.

And of course at the intersection of the glib and the trite is the misquote. Twitter’s favourite is probably this:

In the 9 hours since midnight today, this has been repeated and amplified 20 times. It’s an admirable sentiment, but as Brian Morton pointed out in the New York Times in 2011, Gandhi didn’t say it. If you care about accuracy, there’s no excuse for not discovering that when doing your fact checking … unless of course you don’t care about accuracy and you don’t do any fact checking.

I do hope that I’m wrong about this. I would love Twitter to be a friendly exchange of useful information, banter and conversation, and indeed, I continue to read interesting, sometimes challenging, Tweets from the more thoughtful people I follow.

However, lately Twitter seems to have become more of a haven for the lazy and the inaccurate, not to mention the hateful. I hope that in the run up the UK election we’ll manage to have some better informed debate than this. I somehow doubt we will.

The L&D skills gap on #Ozlearn chat


I’m very much looking forward to joining Con Sotidis for an #Ozlearn tweet chat on Tuesday 10th February at 8 pm Melbourne time, 9 am London time. Together, we’ll be discussing something that’s become increasingly important over the past few years: the L&D skills gap.

L&D, like many professions, faces dramatic changes today, changes with their roots in technological, economic and geopolitical shifts that began in around 1990. Like most professions, however, we are still largely equipped with structures, attitudes and skills sets suitable for the world as it existed before that date.

So just what is our current L&D skills gap? How short are we of the skills we need to be effective in the new, connected, global age? For an answer, I’ve looked at data from the LPI Learning Capability Map. The Capability Map is a description of the 27 skills of the L&D profession, arranged over 4 levels. Nobody suggests you need all 27 of them individually, but a department should certainly have access to them all, whether internally or via partners. (You can self-assess against the LPI Learning Capability Map here: It’s free.)

Just 6 months after the Capability Map was made public, we published an initial report on the first 980 people who had assessed themselves against it (click to download the report). We are now preparing a 30-month report for March this year, based on some 2,000 self-assessments.

The results from the 6- and the 30-month analysis are entirely consistent and show that – perhaps unsurprisingly – the highest ranked skills are those in areas where L&D has traditionally operated, and continues to do most of its work, such as face-to-face delivery and content design. There are strong signs, however, that the skill of virtual/online delivery of courses is growing far faster than these traditional skills.

In contrast, some of L&D’s weakest skills are in the newer, but essential, areas of social learning and developing collaborative skills in others. Hearteningly, however, this also appears to be where we’re developing our skills fastest.

‘Communication, marketing and relationship management’ remains among the seven least popular skills, and yet to my mind is now essential for every L&D department and professional. The reason: it’s core to performance consulting – an essential facet of L&D today.

I cannot pre-empt the entire 30-month report here in the limits of a blog entry, but I am looking forward to exploring this topic in more detail with Con and his #Ozlearn cohort on Tuesday. Here are the questions that I’d like us to reflect on for our hour together:

1) What skills do you feel L&D needs for the future?
2) How do you know where you and your L&D department skills are right now?
3) Where do you see classroom delivery in 3 years? Will it be online delivery, e-learning, self-access resources or something else? Whatever it is – are you ready?
4) Does your organization management, workforce and L&D department view social learning the same way?
5) How do you currently handle communication, marketing and relationship management with the rest of your organisation?

Con, of course, is an expert at running these sessions and will provide the precise wording of these questions. If you’re in the right time zone, please do come and join us for a discussion about what I believe is an essential topic for everyone in L&D today.

An end-of-year wish for L&D

As we move towards the end of the year, many of us will want to take some time out. It seems like it’s been another year of too much to do and not enough time to do it. This sense of being overwhelmed is widespread among everyone I talk to in our field, in every sector and every location.

And it’s based in reality. L&D genuinely is under pressure to do more, usually with fewer resources.

There’s an issue with this pressure, and it applies to everyone who experiences it, whatever field they work in. Abraham Lincoln reputedly said “If I had a tree to feel and eight hours to do it. I’d spend six hours sharpening my ax and two hours chopping.” Whether he actually said this or not, the point stands: we work best when our tools – and our brains – are sharp.

Conversely, when we work flat out, our tools lose their edge, anything achieving becomes harder, and takes longer, and we slip into a downward spiral of decreasing effectiveness.

What we’re left with at the end of this is a learning and development profession that has no time to either learn or develop. That is bad for us, and bad for the organisations and people we serve.

So, if this year it felt like you haven’t had time to pause and take a breath this year, perhaps it would be a good idea to use any holiday time at the year’s end to accomplish two things.

First, take a break. Do not read anything about L&D, design, corporate strategy or anything to do with work. Concentrate instead on unbending from the working year, stretching muscles you haven’t used for a while. And have a bit of fun too.

The second thing I’d ask is this: make a new year’s resolution. Now these are notoriously tricky things to make work, but aim to keep at this one for at least 4 weeks. If you can make it to the end of the first month of the year, the odds are you’ll carry on successfully to the end.

The resolution I’d ask you to make is this: make time to read a good book in our field. Reading helps lift your head from the remorseless grindstone of the week’s daily chores, and gives you time to reflect. It is a slow, deep process that switches your mind from the immediate to the longer-term. And with this unbending, comes an additional benefit. There’s a fair chance you’ll learn something new, and that your axe will end up a little sharper.

Happy reading.

This, my final blog entry for the year, is the preface from the December 2014 edition of Inside Learning Technologies and Skills Magazine. Click to read.