Your ceiling, not your speed
You have automatic outlines for 0% of running English. That is what caps you — not your hand.
How to read the claims in this course
Every load-bearing claim carries a marker. Numbered ones key to the source list, so you can check any single sentence rather than trusting the whole document.
- Sourced fact traceable to a numbered source
- Inference my reasoning or a model — in no source
- Vendor framing true, but written by someone selling it
- Contested the evidence genuinely disagrees
- Single study real, but unreplicated
What this course is not
- It is not a Teeline textbook, and it does not contain the alphabet. The letterforms are lookup-able facts and I will not reproduce them from memory — that is exactly the kind of detail that gets confidently invented. You need a real chart before Week 1. See Before you start.
- It will not get you to 100 wpm. It gets you a correct system, a readable hand at 60, and a regimen. The remaining 40 wpm is drill hours that no course design shortcuts. A UK journalism degree budgets 150 learning hours for a minimum pass at 80 wpm — in the second year of shorthand Sourced fact S3.
- It does not cover NCTJ exam technique, court or medical vocabulary, or Teeline's advanced special outlines.
- It does not claim shorthand will make you understand meetings better. The famous study on handwritten notes argued the opposite — that verbatim transcription is shallower processing than writing in your own words Single study S10. Its own direct replication then found no consistent difference between any group, including people who took no notes at all Contested S11. So: this course buys you a better record. Any claim it buys you better comprehension is unsupported in both directions.
Before you start
The one dependency this course does not satisfy
Teeline's letters are derived from the printed alphabet with the redundant strokes stripped away, which is why it is the fastest of the major systems to learn Sourced fact S5. That derivation is a principle I can teach. The twenty-six specific shapes are not — get them from a real chart, not from me.
Either of these works:
| Source | Cost | Covers |
|---|---|---|
| teeline.online | free | alphabet, early lessons, worked outlines |
| Teeline Gold Standard for Journalists, Cartwright | ~£30 | the whole system to 100 wpm; the book the NCTJ builds its distance course around Sourced fact S1 |
You also need: a pen you can write with for twenty minutes without gripping, a lined pad, and a way to play speech at a controlled tempo. That last one is not optional — it is the whole engine of Weeks 4–6, and it is the single practical reason Teeline beats Gregg for a solo learner. The NCTJ's own distance plan is built around graded audio running 60 to 100 wpm Sourced fact S2.
Where the hours go
| Raw estimate | Published | |
|---|---|---|
| Six weeks of blocks | ~12 hrs | ~30 hrs |
| After the course, 60 → 100 | — | 100–200 hrs |
The published figure is the raw estimate multiplied by 2.5, because self-study learners consistently overrun their own estimates by roughly that factor Inference. If 30 hours across six weeks is not real for you, the honest move is to decide that now rather than in Week 3.
WEEK 01
Write something unreadable, today
Threshold 1 opens: you are not writing words. You are writing a skeleton that a specific reader — you, tomorrow — will reconstruct.
Most shorthand courses open with the alphabet and reach real dictation around week five. That ordering is why people quit in week two. You are doing the entire loop — hear, encode, wait, decode — in the first sitting, badly, so that the thing you improve is a thing you have already done.
Here is the idea the whole system rests on. Longhand is a lossless encoding: every letter is present, and anyone can read it, at any distance in time. Shorthand is lossy, deliberately, and it is tuned for one decoder in one context — you, who was in the room, reading within a day. Teeline drops vowels that are not the first or last letter of a word Sourced fact S5 not because vowels are unimportant, but because you already have them from context, and the pen does not.
This is why shorthand rots. A page you can read in an hour may be genuinely unrecoverable in a month — you were the decoder, and the context you were decoding against is gone.
Read this week's section. Write down, in one sentence, what you expect to be hardest. Keep it — Week 5 asks you to check it.
Get the alphabet chart. Copy each letter twenty times, saying the letter aloud as you write it. Do not try to memorise — that comes from use. You are building a motor habit, not a lookup table.
Have something read two sentences to you slowly — roughly 30 wpm, half of comfortable speech. Write them in Teeline. Look up letters mid-sentence. It will be terrible. Do not rewrite it neatly.
Twenty-four hours later, transcribe what you wrote, without the audio. Mark every word you could not recover. That list is your actual syllabus.
For each unrecovered word, write why: wrong stroke, missing stroke, or right stroke and no context. The three failures have three different fixes, and telling them apart is most of the skill.
You wrote an outline for "management" and tomorrow you cannot read it. Before you look at anything: which of the three failures is most likely, and what does that imply about the fix?
Right stroke, no context — almost always, for a word like this. "Management" is a content word carrying most of the sentence's meaning, so unlike "the" or "of" you cannot reconstruct it from the grammar around it. Its skeleton after vowel elision is thin, and thin skeletons for content words are where lossy encoding actually loses.
The fix is therefore not "write more carefully." It is that high-value, low-redundancy words need a fuller outline than the general rule gives them — you spend strokes where context cannot pay you back. Notice that this is the opposite of the trade you make on function words, and hold onto that: Week 3 turns it into the course's central economy.
WEEK 02
The skeleton
Threshold 1 crosses: stop transcribing, start encoding.
Last week you looked up letters while writing. This week you will not, which means you will write wrong outlines instead. That is the correct trade and it is not a compromise — a wrong outline you can reconstruct is worth more than a right outline that cost you three seconds, because the three seconds cost you the next clause.
You are supposed to feel stuck here
Week 2 is where most people conclude they are bad at this. The specific sensation: you know the letters, and you still cannot produce them at speed without stopping. That gap is not a sign you learned the alphabet badly. It is the gap between knowing a thing and having automated it, and it closes with hours, not insight. If Week 2 feels like Week 1 with more frustration, you are exactly where you should be.
Four short sessions, not one long one. Copy connected prose — not word lists — from anything to hand. Never stop to look a letter up. If you do not know it, invent something and circle it.
A passage at 40 wpm. Same rule: no lookups, ever, for any reason.
Transcribe cold. Then check every circled invention against the chart.
Answer in prose, for yourself: which words did you lose, and did they share anything? Most people find their losses cluster. Say what the cluster is.
You are taking dictation and hit the phrase "we should probably review the department's budget." Three of those words deserve very different treatment. Name which words get which treatment, and why — do not describe the outlines, describe the decisions.
"we", "the" — maximum compression, no thought. These are so frequent and so recoverable from grammar that any stroke beyond the minimum is waste. They should cost you nothing because they should not be decisions at all.
"probably" — the interesting case. Long, but low-information: drop it entirely and the sentence survives with its meaning almost intact. The honest treatment is a stub, or nothing, and this is the first time you will feel the difference between a transcript and a record.
"budget" — spend here. Short word, but it is the only thing in the sentence that cannot be reconstructed from anything else. Content words in a domain you know are the exception to compression.
If you answered in terms of stroke shapes rather than decisions, reread Week 1's close. The system is the easy half; knowing where to spend is the skill.
WEEK 03
Where the words actually are
Threshold 2: frequency is the entire economy. A hundred outlines buy you half of English.
Here is the fact the whole craft is built on, and it is much more extreme than it sounds. In a ten-million-word sample of the Cambridge English Corpus, the top ten words cover about a quarter of everything you will ever encounter, and the top hundred cover close to half Sourced fact S7. An independent count on the Wall Street Journal treebank lands in the same place: the hundred most frequent words are 48.1% of all tokens, and the top five hundred are 63.6% Sourced fact S8. Two unrelated corpora, two methods, the same answer.
Read that as an investment schedule. A hundred outlines — a week of work — automates half of everything anyone will ever say to you. The next four hundred buy fifteen points. The five hundred after that buy about eleven Sourced fact S9. Everything past roughly a thousand is the long tail, and no amount of study makes it pay.
Which means: the most common mistake in learning shorthand is studying it evenly. The brief-form list is not a vocabulary to get through. It is a ranked buy order, and the first hundred entries matter more than the rest of the book combined.
Find the short-form list in your source. Rank it against a frequency list of English rather than taking the book's order.
The top hundred only. Spaced across five days, not massed — distributed practice is one of the two best-evidenced study techniques there is Sourced fact S15. Write each in context, in a sentence, never as an isolated item.
A 50 wpm passage. Count how many times you hesitated. Write the number down — Week 5 needs it.
Week 1 claimed shorthand "rots" — readable in an hour, unrecoverable in a month. You now know the frequency structure of English. Explain the rot in terms of it. Why does a page of shorthand decay unevenly?
The frequent half of your page never rots. Function words are recoverable from grammar alone, forever, by anyone — they carry almost no information, which is exactly why they compress so well. The decay is entirely in the long tail: the proper nouns, the figures, the domain terms, the one number someone said once. Those are the words with no redundancy to reconstruct from, and they are precisely the words you took the note for.
So shorthand decay is not uniform fading. It is the selective loss of exactly the content you cared about, while the connective tissue stays perfectly legible. This is why readback discipline is a same-day habit and not a preference, and it is also why every week of this course puts the readback twenty-four hours later rather than immediately — an immediate readback measures nothing, because nothing has decayed yet.
WEEK 04
Joining
Threshold 2 extends: the unit of writing is not the word.
Teeline's practical ceiling comes largely from a habit rather than from the system: experienced writers invent their own groupings, joining common word sequences into single unbroken outlines, and that is a substantial part of how speeds near 150 wpm get reached at all Sourced fact S5. Every pen lift costs time. A phrase written as one movement is not four outlines quickly — it is one outline.
This is also where the course stops being about Teeline and starts being about you. The highest-value groupings are not in any book, because they are the phrases you hear twenty times a week. Someone taking board minutes and someone taking clinical notes should end up with visibly different shorthand, and if yours looks like the textbook's after a year, you have not finished learning it.
Go through your last month of real notes. Find the phrases that recur. Not general English — your English. Names, recurring constructions, the four things people always say in your meetings.
Design a joined outline for each. Test: can you write it in one movement, and read it back tomorrow with nothing else on the page? If either fails, it is a bad grouping. Discard it and do not be precious — most first attempts fail.
Your groupings only, in running prose.
55 wpm. Count hesitations again.
Four situations. For each, name which of the course's principles applies — vowel elision, short form, word grouping, or spend-more-strokes — and say why the other three are wrong.
(a) A name you will hear forty times this quarter.
(b) "In my opinion."
(c) A dollar figure, said once.
(d) "Organisation."
(a) Personal short form. Frequency is local, not global. This name is nowhere near the top thousand words of English but it is in your top hundred, and the frequency argument applies to the corpus you personally face. Not a grouping — it is one word.
(b) Word grouping. Four words, fully predictable in sequence, near-zero information. One movement. Elision alone would still cost you four outlines and three pen lifts.
(c) Spend more strokes. The instinct after three weeks of compression is to compress it, and that instinct is wrong here. Zero redundancy, zero recoverability, said once, and almost certainly the reason the note exists. Write it fully. Consider writing it twice.
(d) Vowel elision, and nothing clever. Long, but ordinary — the general rule handles it. Not frequent enough in most people's speech to earn a short form, not part of a fixed phrase. The skill here is recognising the case that needs no special treatment, which is the hardest one to see once you have learned four techniques and want to use them.
WEEK 05
The hesitation audit
Threshold 3: speed is not hand speed. Speed is the absence of decisions.
Take the numbers you have been writing down. At 60 wpm you have one second per word. An outline you know costs perhaps four-tenths of that. An outline you have to think about costs a second or more — and worse, you lose the words spoken during the thinking, so a single hesitation typically costs two or three words rather than one.
Run the arithmetic and the conclusion is uncomfortable: writing faster is nearly worthless, and eliminating a hesitation is worth roughly three times what it looks like Inference. The meter at the top of this page is built on exactly that model — it computes a ceiling from your coverage, and you can inspect the equation. It is a teaching device, not a measurement of you.
The Week 5 dip is real
When you start consciously auditing your pauses, your speed usually drops for a week or two. You have made an automatic process deliberate in order to inspect it, which is the only way to fix it and is also temporarily disabling. Push through this by continuing to take dictation, not by returning to what felt smooth. Difficulty that slows you down now and improves retention later is the mechanism, not a side effect Sourced fact S14.
Take a 55 wpm passage and mark the page every time you pause — a tick in the margin, no analysis. Then transcribe and list what caused each one.
Sort by cost, which is frequency × delay, not delay alone. A half-second pause on a word you hit thirty times beats a two-second pause on a word you hit once.
Your top twenty only. This is the highest-return practice in the entire course, and it is available to you only because you measured.
Find the Week 1 note where you predicted what would be hardest. Compare. Most people are wrong about themselves in a specific and useful direction.
Someone tells you: "I've plateaued at 70 wpm, so I'm doing timed drills every day — same passage, pushing to write faster and faster until my hand can keep up."
Everything in that plan is wrong. Identify each error and what it should be instead.
"Until my hand can keep up." The hand is not the constraint and never was. Pen velocity is a rounding error next to lookup latency. They are training the one variable that does not bind.
"Same passage." Repeating one passage trains recall of that passage. Within three days they are writing from memory, the hesitations disappear, and the measurement becomes meaningless while feeling like progress — fluency mistaken for learning, which is the defining failure of solo study.
"Every day, pushing faster." No diagnosis anywhere. They have not asked which words cost them the plateau. Without the audit they are practising the 95% they already have automatic, at the expense of the 5% that is actually capping them.
What it should be: varied unseen passages at a speed they can just sustain accurately; hesitations marked and ranked; targeted drill on the top twenty; then a tempo increase of five wpm only once accuracy holds. Speed is a result you are permitted to have, not a thing you can practise directly.
WEEK 06
Build your own ladder
All three thresholds, applied without scaffolding. This week has no worked example.
Weeks 1 and 2 showed you fully worked answers. Weeks 3 and 4 gave you partial ones. This week gives you none, because the thing being tested is whether you can now diagnose yourself — and the honest measure of this course is not your speed in Week 6 but whether you are still practising in month six. That is where shorthand is actually lost: a working journalist who passed at 100 wpm reported losing substantial speed within months of stopping Vendor framing S13 — a practitioner account in a piece promoting shorthand, so read the direction as reliable and the magnitude as not.
An unseen 60 wpm passage. Transcribe at 24 hours. Score it: words lost, and which of the three failure types each was.
Write your regimen. It must specify: session length, weekly frequency, where unseen passages come from, how you decide to raise the tempo, and what you do when you miss a week. No template — if you cannot write it yourself you have not crossed Threshold 3.
Put it in your calendar as recurring blocks, and tell one person you are doing it. Cohort formats complete at substantially higher rates than solo self-paced study, and isolation is among the most-cited reasons people abandon courses Sourced fact S15. One other person is a cheap intervention on the highest-risk variable.
Predict your accurate, sustainable speed on an unseen passage in six months, assuming you run the regimen you just wrote. Then say what evidence you are using — and whether it is evidence about shorthand or about yourself.
There is no answer here. The item exists so that in six months you have a dated, specific prediction to check yourself against — which is the only calibration instrument available to someone studying alone.
One thing worth knowing before you write the number: people reliably predict their future knowledge will stay roughly where it is now, discounting both further study and forgetting. Whatever you wrote is probably too close to today's speed in both directions.
AFTER
60 → 100 → 150
The course ends at 60 wpm and a regimen. The rest is a longer road than most people are told, so here it is plainly.
| Stage | Rough hours | What changes |
|---|---|---|
| 60 → 80 | 40–80 | Short forms become automatic. Mostly drill. |
| 80 → 100 | 60–120 | Your own groupings do the work. This is the long middle. |
| 100 → 150 | not hours — years | Ceiling territory. Teeline reaches it, but few writers do. |
Those figures are my estimates, not measurements Inference, anchored on one hard published number: a second-year university Teeline module allocates 150 total learning hours to reach a minimum pass at 80 wpm Sourced fact S3. On 100, the target is well established — it is the NCTJ's gold standard and the threshold for professional eligibility Sourced fact S1, S6. On 150, be careful what you read: it is described as achievable in Teeline largely through personal word groupings Sourced fact S5, but the sources making that claim are promotional rather than measured Vendor framing. Treat 150 as the top of the range, not the plan.
SOURCES
Every numbered claim
Marked by how they were checked. Opened means I read the source while building this. Carried means it is established literature I did not re-verify here — treat those as weaker, and check them yourself before relying on them.
- NCTJ, Shorthand — 100 wpm as the professional standard; Cartwright's Teeline Gold Standard for Journalists as the recommended text, designed to take a beginner to 100. nctj.com · Opened
- NCTJ, distance learning — graded audio, 60–100 wpm progression; exams 60–120. nctj.com · Opened
- Staffordshire University, JOUR50277 (Shorthand 2) — 150 total learning hours; minimum pass 80 wpm; second-year module. module descriptor · Opened. The single most useful number in this course, because it is an institution's own budgeting rather than a promise to a prospective student.
- University of Sheffield — 60 then 100 wpm progression. sheffield.ac.uk · Opened
- Teeline shorthand, Wikipedia — alphabet-derived letterforms; vowels dropped when not first or last; 150 wpm attributed to personal word groupings. Opened. Tertiary, and its 150 wpm figure traces to a trade-press citation, so the mechanism is trustworthy and the number is not.
- teeline.online — 60 for the diploma, 100 gold standard, some writers past 120. teeline.online · Opened
- New General Service List, Coverage — 10-million-word Cambridge English Corpus sample: top 10 words ≈25%, top 100 ≈50%, top 2000 = 83%. newgeneralservicelist.com · Opened
- Penn Treebank / Wall Street Journal counts — top 10 slightly over 25%; top 100 = 48.1%; top 500 = 63.6%. Opened. Deliberately included as an independent corpus and method from S7, because a single frequency source would be the weakest link in the course's central argument. They agree.
- Nation & Waring (1997), vocabulary size and text coverage, Brown corpus (Francis & Kučera 1982). lextutor.ca · Opened
- Mueller & Oppenheimer (2014), The Pen Is Mightier Than the Keyboard, Psychological Science 25(6), 1159–1168 — verbatim transcription as shallower processing. Opened (abstract). Famous, and the direct ancestor of nearly every "write your notes by hand" claim you have encountered.
- Morehead, Dunlosky & Rawson (2019), Educational Psychology Review 31(3), 753–780 — direct replication and extension of S10. No consistent differences between groups, including a group that took no notes; meta-analysis of direct replications found small non-significant effects favouring longhand. Springer · Opened (abstract). I searched for this specifically rather than citing S10 alone, and it substantially changed the non-goals of this course.
- Practitioner account, trainee reporter on losing speed after passing at 100 wpm, in a piece promoting shorthand learning. Opened. Tier 4 — direction only.
- Bjork & Bjork (2011), Making Things Hard on Yourself, But in a Good Way — desirable difficulties; the learning/performance distinction. Carried
- Hattie & Donoghue (2021), meta-analysis of ten learning techniques; Dunlosky et al. (2013) — distributed practice and practice testing as the two best-supported techniques. Cohort completion and isolation effects from MOOC attrition literature. Carried
S12 is deliberately absent — it was a forum post about Gregg's memory load that informed the system choice but supports nothing in the course itself, and leaving the gap visible is cheaper than renumbering.