Who said it

Each delegation measured against its own record: what share of its own speeches used this vocabulary, rather than how often it turns up in the corpus overall. Of 601 speakers, 133 spoke often enough for that share to mean anything.

Speakers by rate

Which delegations used the vocabulary most, as a share of their own speeches?

Measure Period Ranked by

Map

Loading the background map…

The radius of a circle carries the rate, not its area. Read the table for the numbers.

How to read this

One circle per delegation that spoke often enough to be measured. The radius carries the same figure the table is ranked by; colour carries nothing. A heavier ring marks a delegation whose three-letter country code (ISO3) another speaker shares.

Click a circle to pick out its row in the table, or a row to pick out its circle. Two country codes are held by more than one speaker and are marked with an asterisk in the table; in every period shown, only one holder of each has spoken often enough to be measured, so nothing on this map stands for two delegations at once. Circles that would land on the same coordinates are grouped rather than stacked, so that stays true as the corpus grows.

What it does not show

A country's position on this map is a way of finding it, nothing more. It does not locate the speaker: every speech in this corpus was delivered in the Security Council chamber in New York.

468 speakers delivered fewer than 100 speeches in this period and carry no rate. They are not ranked low; they are not ranked. A speaker gets no rate for any period in which it delivered fewer than 100 speeches. That threshold is the point at which a zero starts to mean something: across the corpus as a whole, 3.08% of speeches use this vocabulary, so seeing none of it in fewer than 96 speeches is exactly what the Council average would predict. The counts are published either way, because a count is a fact and a rate is an estimate.

The table behind the map

The same 133 rows, in the same order. The table is the main presentation and the map is a way into it: a circle cannot be reached by keyboard or read aloud by a screen reader, and a row can.

Speakers ranked by share of its speeches for genocide, 1992-2023
Speaker	Group	Speeches	Using the term	Share	Per 100,000 words
	Eastern European Group	124	35	28.23%	100.53
	African Group	697	187	26.83%	243.74
	Western European and Others Group	217	46	21.20%	35.06
	Eastern European Group	404	56	13.86%	30.81
	Civil society	135	18	13.33%	27.41
COD*	African Group	173	23	13.29%	14.74
	Asia-Pacific Group	246	32	13.01%	16.79
	Asia-Pacific Group	150	19	12.67%	17.39
SRB*	Eastern European Group	177	19	10.73%	13.54
	African Group	236	23	9.75%	17.61
	Eastern European Group	308	29	9.42%	20.60
	Eastern European Group	461	43	9.33%	20.72
	Latin American and Caribbean Group	233	19	8.15%	12.64
	Eastern European Group	478	38	7.95%	18.01
	Western European and Others Group	430	33	7.67%	17.11
	Western European and Others Group	388	28	7.22%	25.72
	African Group	109	7	6.42%	15.29
	Western European and Others Group	686	44	6.41%	12.72
	Asia-Pacific Group	191	12	6.28%	10.87
	African Group	319	20	6.27%	11.60
	African Group	266	16	6.02%	12.68
	Intergovernmental organisation	882	51	5.78%	6.89
	Asia-Pacific Group	457	26	5.69%	10.91
	Eastern European Group	141	8	5.67%	10.72
	Asia-Pacific Group	602	34	5.65%	10.63
	Latin American and Caribbean Group	448	25	5.58%	11.28
	Western European and Others Group	983	54	5.49%	10.61
	Latin American and Caribbean Group	680	37	5.44%	13.70
	Latin American and Caribbean Group	166	9	5.42%	14.57
	Western European and Others Group	334	18	5.39%	11.07
	Western European and Others Group	207	11	5.31%	33.11
	State	248	13	5.24%	5.15
	Eastern European Group	560	27	4.82%	23.97
	United Nations	4,709	227	4.82%	7.73
	Latin American and Caribbean Group	297	14	4.71%	8.72
	African Group	439	20	4.56%	6.24
	African Group	510	23	4.51%	9.36
	Latin American and Caribbean Group	752	33	4.39%	9.69
	Asia-Pacific Group	712	31	4.35%	5.94
	Western European and Others Group	332	14	4.22%	16.01
	Asia-Pacific Group	216	9	4.17%	8.59
	Eastern European Group	962	40	4.16%	15.49
	African Group	196	8	4.08%	7.10
	Latin American and Caribbean Group	1,329	54	4.06%	12.10
	African Group	326	13	3.99%	7.52
	Latin American and Caribbean Group	234	9	3.85%	8.17
	Western European and Others Group	694	26	3.75%	9.17
	Asia-Pacific Group	1,205	45	3.73%	10.02
	Asia-Pacific Group	408	15	3.68%	6.80
	Western European and Others Group	867	31	3.58%	9.96
	African Group	592	21	3.55%	6.40
	African Group	347	12	3.46%	22.11
	African Group	586	20	3.41%	5.99
	African Group	823	27	3.28%	7.63
	African Group	122	4	3.28%	9.03
	African Group	590	19	3.22%	7.74
	Western European and Others Group	373	12	3.22%	8.10
	Western European and Others Group	995	32	3.22%	9.58
	Eastern European Group	591	19	3.21%	9.61
	Western European and Others Group	596	19	3.19%	6.54
	Asia-Pacific Group	416	13	3.13%	7.05
	Asia-Pacific Group	740	22	2.97%	6.34
	Western European and Others Group	5,484	163	2.97%	10.92
	African Group	242	7	2.89%	7.30
	Latin American and Caribbean Group	1,102	31	2.81%	9.23
	Asia-Pacific Group	143	4	2.80%	2.30
	Asia-Pacific Group	914	25	2.74%	8.98
	Asia-Pacific Group	189	5	2.65%	11.80
	Intergovernmental organisation	266	7	2.63%	2.39
	Western European and Others Group	838	22	2.63%	11.42
	Western European and Others Group	4,973	127	2.55%	7.79
	Western European and Others Group	825	21	2.55%	5.41
	Latin American and Caribbean Group	1,412	35	2.48%	5.67
	Latin American and Caribbean Group	450	11	2.44%	4.65
	Western European and Others Group	371	9	2.43%	4.78
	Western European and Others Group	793	19	2.40%	8.75
	Latin American and Caribbean Group	628	15	2.39%	7.29
	Western European and Others Group	429	10	2.33%	3.74
	African Group	258	6	2.33%	3.79
	Asia-Pacific Group	774	18	2.33%	5.60
	African Group	529	12	2.27%	4.81
	African Group	404	9	2.23%	5.51
	Asia-Pacific Group	676	15	2.22%	6.82
	Asia-Pacific Group	407	9	2.21%	6.77
	Western European and Others Group	922	20	2.17%	5.31
	African Group	659	14	2.12%	6.30
	Latin American and Caribbean Group	962	20	2.08%	8.81
	Western European and Others Group	338	7	2.07%	5.00
	African Group	1,114	23	2.06%	9.35
	Latin American and Caribbean Group	100	2	2.00%	1.69
	African Group	603	12	1.99%	3.99
	Asia-Pacific Group	102	2	1.96%	2.60
	African Group	529	10	1.89%	6.93
	African Group	437	8	1.83%	7.00
	Latin American and Caribbean Group	385	7	1.82%	6.58
	Western European and Others Group	1,928	35	1.82%	5.74
	Asia-Pacific Group	944	17	1.80%	4.22
	Western European and Others Group	5,518	94	1.70%	6.29
	Western European and Others Group	1,352	23	1.70%	6.83
	Latin American and Caribbean Group	550	9	1.64%	3.47
	Western European and Others Group	688	11	1.60%	3.12
	Western European and Others Group	1,001	16	1.60%	5.26
	African Group	320	5	1.56%	4.38
	African Group	1,347	21	1.56%	2.91
	African Group	257	4	1.56%	4.95
	African Group	974	15	1.54%	2.59
	African Group	522	8	1.53%	3.04
	African Group	463	7	1.51%	6.33
	Western European and Others Group	889	13	1.46%	4.16
	Western European and Others Group	276	4	1.45%	4.91
	Latin American and Caribbean Group	856	12	1.40%	2.74
	Latin American and Caribbean Group	753	10	1.33%	3.54
	African Group	948	12	1.27%	5.54
	Western European and Others Group	1,121	14	1.25%	3.40
	Asia-Pacific Group	1,208	15	1.24%	3.13
	Latin American and Caribbean Group	1,937	23	1.19%	3.01
	Asia-Pacific Group	1,371	16	1.17%	2.75
	Western European and Others Group	449	4	0.89%	2.65
	Asia-Pacific Group	471	4	0.85%	1.55
	Western European and Others Group	5,101	43	0.84%	2.30
	African Group	362	3	0.83%	2.44
	Asia-Pacific Group	137	1	0.73%	1.93
	African Group	148	1	0.68%	0.79
	African Group	334	2	0.60%	0.87
	African Group	351	2	0.57%	3.16
	African Group	180	1	0.56%	1.17
	Asia-Pacific Group	2,055	11	0.54%	1.16
	Asia-Pacific Group	585	3	0.51%	1.13
	Asia-Pacific Group	4,737	20	0.42%	1.37
	Asia-Pacific Group	976	4	0.41%	0.91
	African Group	319	1	0.31%	0.61
	African Group	127	0	0.00%	0.00
	African Group	130	0	0.00%	0.00

Who held a seat when they spoke

Was a delegation speaking as a member of the Council, or from outside it?

Period Showing Ordered by seated in some years, not in others

P5seatE10seatNon-member stateUNNon-state

Speakers by the positions they held when speaking, 1992-2023
Speaker	Speeches	From a seat	E10	Non-member state
Japan	2,055	77.96%	1,602	453
Brazil	1,937	82.55%	1,599	338
Germany	1,928	86.62%	1,670	258
Mexico	1,412	87.68%	1,238	174
India	1,371	72.72%	997	374
Spain	1,352	88.54%	1,197	155
South Africa	1,347	77.06%	1,038	309
Argentina	1,329	81.34%	1,081	248
Indonesia	1,208	73.76%	891	317
Pakistan	1,205	68.46%	825	380
Norway	1,121	75.65%	848	273
Nigeria	1,114	84.47%	941	173
Peru	1,102	85.12%	938	164
Poland	1,001	84.92%	850	151
Belgium	995	84.42%	840	155
Ukraine	983	64.60%	635	348
Viet Nam	976	89.86%	877	99
Egypt	974	61.40%	598	376
Albania	962	92.31%	888	74
Chile	962	83.16%	800	162
Gabon	948	98.84%	937	11
United Arab Emirates	944	85.49%	807	137
Italy	922	71.80%	662	260
Malaysia	914	73.30%	670	244
Sweden	889	88.19%	784	105
Netherlands	867	78.66%	682	185
Colombia	856	73.48%	629	227
New Zealand	838	76.73%	643	195
Ireland	825	85.33%	704	121
Ghana	823	94.90%	781	42
Turkey	793	43.38%	344	449
South Korea	774	61.89%	479	295
Bolivia (Plurinational State Of)	753	96.15%	724	29
Uruguay	752	82.71%	622	130
Kuwait	740	85.41%	632	108
Syria	712	32.44%	231	481
Rwanda	697	75.18%	524	173
Australia	694	53.75%	373	321
Portugal	688	71.66%	493	195
Canada	686	42.27%	290	396
Venezuela (Bolivarian Republic Of)	680	74.26%	505	175
Jordan	676	72.78%	492	184
Kenya	659	82.55%	544	115
Ecuador	628	78.03%	490	138
Cote D'ivoire	603	83.91%	506	97
Bangladesh	602	57.97%	349	253
Switzerland	596	61.58%	367	229
Tunisia	592	79.39%	470	122
Estonia	591	82.57%	488	103
Angola	590	85.93%	507	83
Morocco	586	50.00%	293	293
Kazakhstan	585	75.73%	443	142
Lithuania	560	88.75%	497	63
Dominican Republic	550	92.91%	511	39
Ethiopia	529	79.40%	420	109
Uganda	529	82.99%	439	90
Equatorial Guinea	522	98.66%	515	7
Senegal	510	81.96%	418	92
Azerbaijan	478	71.34%	341	137
Lebanon	471	59.87%	282	189
Niger	463	97.19%	450	13
Croatia	461	60.95%	281	180
Guatemala	450	55.78%	251	199
Malta	449	85.97%	386	63
Costa Rica	448	69.20%	310	138
Libya	439	61.05%	268	171
Mozambique	437	89.24%	390	47
Austria	429	71.79%	308	121
Singapore	416	84.86%	353	63
Qatar	408	52.70%	215	193
Philippines	407	63.88%	260	147
Bosnia and Herzegovina	404	51.49%	208	196
Namibia	404	75.00%	303	101
Luxembourg	388	81.19%	315	73
Jamaica	385	91.95%	354	31
Denmark	373	77.48%	289	84
Romania	371	76.82%	285	86
Guinea	362	93.65%	339	23
Mali	351	72.93%	256	95
Chad	347	92.22%	320	27
Greece	338	77.22%	261	77
Cameroon	334	97.90%	327	7
Slovak Republic	334	58.98%	197	137
Hungary	332	77.71%	258	74
Algeria	326	52.45%	171	155
Burkina Faso	320	90.94%	291	29
Togo	319	96.24%	307	12
Slovenia	308	54.87%	169	139
Saint Vincent And The Grenadines	297	97.31%	289	8
Bulgaria	276	84.42%	233	43
Botswana	266	71.43%	190	76
Mauritius	258	96.12%	248	10
Benin	257	84.44%	217	40
Congo	242	92.15%	223	19
Tanzania	236	71.19%	168	68
Panama	234	82.91%	194	40
Bahrain	216	54.17%	117	99
Czech Republic	207	79.23%	164	43
Djibouti	196	71.43%	140	56
Oman	189	87.30%	165	24
Gambia	180	91.67%	165	15
Honduras	166	84.94%	141	25
Guinea-Bissau	148	70.27%	104	44
Cape Verde	127	91.34%	116	11
Zimbabwe	99	39.39%	39	60

What the five positions mean

The UN Charter provides two kinds of Council membership: five permanent members (P5) and ten elected ones (E10). A speech counts as 'seated' when its speaker held one of those places in the year it was delivered. The other three groups are not a single category: a non-member state was simply not on the Council that year, the UN Secretariat never can be, and a non-state speaker was invited to address it. All five counts are published separately so that distinction survives, and together they add up to every speech that speaker gave.

A row is a mixture, not a single label. Width is the share of that speaker's own speeches.

How to read this

Each row is one speaker's own speeches, divided up by the position it held when it gave them. The two seated bands are the UN Charter's two kinds of Council membership: the five permanent members and the ten elected ones. The three unseated bands are kept apart because they are not the same thing as each other.

Of 601 speakers in this period, 105 spoke both from a seat and from outside one, 5 only ever from a seat, and 491 never from one. Japan is the clearest case: 2,055 speeches, 77.96% of them from a seat.

What it does not show

Membership belongs to a speech rather than to a country. The ten elected seats rotate, so no speaker has a single fixed status: Rwanda spoke as an elected member in 1994 and again in 2013-2014, and as a non-member in every other year of the corpus. Each row here is therefore a mixture rather than a label, and shading a speaker in one colour would erase that change. Unlike the rates elsewhere on this page, these figures are published however few speeches lie behind them: a share of a speaker's own known speeches is a fact about the record rather than an estimate drawn from it.

There is no minimum sample here, unlike every rate on this page. The UN Charter provides two kinds of Council membership: five permanent members (P5) and ten elected ones (E10). A speech counts as 'seated' when its speaker held one of those places in the year it was delivered. The other three groups are not a single category: a non-member state was simply not on the Council that year, the UN Secretariat never can be, and a non-state speaker was invited to address it. All five counts are published separately so that distinction survives, and together they add up to every speech that speaker gave.

A whole-corpus row is not the sum of the period rows, and the two should not be read side by side: a speaker can hold a seat for a decade and still count as a non-member across the corpus as a whole.

What a delegation says that the room does not

With the debate held constant, which words set one speaker's language apart from everybody else's?

Speaker Compared against Rows

4,763 paired speeches out of 5,101 (93.37% of its record) · 2,520,462 words against 2,353,382 in the comparison set · pairing reduced the largest differences by 0.39 on the log-ratio scale

Keywords for Russian Federation, against a matched control
Word	Log ratio	This speaker	Compared with	G²
russian*	+1.93 [+1.91, +2.06]	4,850	1,190	2,135
western	+3.48 [+3.25, +3.63]	2,043	171	1,739
kyiv	+4.51 [+4.19, +4.99]	1,410	58	1,456
federation*	+2.27 [+2.13, +2.41]	2,096	407	1,135
situation	+0.89 [+0.87, +0.93]	7,387	3,718	997
believe	+1.33 [+1.33, +1.54]	3,222	1,194	833
convinced	+3.12 [+2.94, +3.34]	985	106	758
settlement	+1.33 [+1.19, +1.33]	2,905	1,083	744
ukrainian	+1.69 [+1.69, +1.92]	1,932	560	708
problems	+1.74 [+1.67, +1.90]	1,832	511	702
colleagues	+1.91 [+1.91, +2.10]	1,606	399	698
attempts	+2.65 [+2.51, +2.78]	946	141	614
unfortunately	+2.25 [+2.09, +2.52]	1,073	210	577
note	+1.29 [+1.22, +1.38]	2,342	895	575
so-called	+2.85 [+2.60, +2.93]	805	104	567
fact	+1.39 [+1.36, +1.52]	2,004	715	551
russia	+0.88 [+0.78, +1.00]	3,791	1,929	496
donbas	+3.58 [+3.40, +4.26]	500	39	436
military	+0.92 [+0.79, +0.92]	2,927	1,440	420
country	+0.71 [+0.65, +0.79]	4,240	2,420	387

2 of the 20 rows drawn are the speaker naming itself.

What it was heard on

The pairing holds the agenda item constant, so this is what was held constant. Russia spoke on 90 items, and 22.43% of its speeches fell on its three commonest.

Israel/Palestine 9.39%
Syria 6.94%
Maintenance Of International Peace And Security 6.10%
Ukraine 5.63%
Afghanistan 5.35%
82 further items 56.50%

Bar length shows the size of the difference, not the confidence in it. Colour shows which way the difference runs.

How to read this

One row per word, ordered by how confident the comparison is (G²). Bar length carries the log ratio: how many times more often this speaker used the word than the speeches it is compared against, doubling with every whole number. A word can reach the top of the table on a small difference, which is why length carries the size of the effect rather than the confidence.

An asterisk marks a word taken from the speaker's own name. Those rows are marked rather than deleted: how often a delegation names itself says something about how it speaks, and removing rows would change the ranking of everything below them.

The bracket after a row is the range its log ratio covered across 10 draws of the comparison set, this one among them. A wide bracket says more about the luck of the draw than about the speaker.

What it does not show

A word appears here because this speaker used it at a different rate from comparable speakers on the same occasion. That is not a position, a policy or a belief, and it says nothing about whether the speaker approved or objected.

Each of the speaker's speeches is paired with one delivered by somebody else in the same year, on the same agenda item and from the same speaker group. No speech is used twice. Where a group holds fewer such speeches than the speaker gave, the shortfall is reported rather than filled from elsewhere, and the coverage figure says how much of the speaker's record the comparison actually rests on.

An asterisk marks a word taken from the speaker's own official name. Every delegation says its own name constantly, and those rows would otherwise fill the top of every table. They are marked rather than removed. The check is mechanical and therefore partial: it catches 'russian' and 'federation' for the Russian Federation, and misses 'french' and 'beijing', which do not appear in the names they belong to. An unmarked row is not a guarantee that the word is not the speaker naming itself.

Who is not here

468 speakers were never paired at all. They delivered fewer than 100 speeches, so no comparison could be built. They appear here as a number rather than a list, because there is nothing to show.
7 were paired and then held back. Unlike the group above, each of these has a figure a reader can weigh, so they stay in the picker and give their reason when selected: Somalia (50.00%),Cape Verde (55.91%),Yemen (50.35%),Haiti (90.00%),Sierra Leone (80.33%),UN (2.61%),European Union (49.77%).
A distinctive word is not a position. A word appears here because this speaker used it at a different rate from comparable speakers on the same occasion. That is not a position, a policy or a belief, and it says nothing about whether the speaker approved or objected. A delegation that names a conflict often may be prosecuting it, deploring it, or chairing the debate about it. Read the record instead

What this table will not tell you

468 speakers have no rate at all. A speaker gets no rate for any period in which it delivered fewer than 100 speeches. That threshold is the point at which a zero starts to mean something: across the corpus as a whole, 3.08% of speeches use this vocabulary, so seeing none of it in fewer than 96 speeches is exactly what the Council average would predict. The counts are published either way, because a count is a fact and a rate is an estimate.
4 of the ranked speakers appear on no map. Civil Society, European Union, UN, African Union are not states and have no place on a globe. They are in the table and missing from the figure above, on purpose.
COD is shared by 2 speakers. Democratic Republic Of The Congo, Zaire carry the same country code, because using a successor state's code is the only way to place a historical state on a map at all. They are never merged: a combined total would belong to no state that ever spoke.
SRB is shared by 3 speakers. Serbia, Serbia And Montenegro, Yugoslavia carry the same country code, because using a successor state's code is the only way to place a historical state on a map at all. They are never merged: a combined total would belong to no state that ever spoke.
A country's map position is not where anyone spoke. A country's position on this map is a way of finding it, nothing more. It does not locate the speaker: every speech in this corpus was delivered in the Security Council chamber in New York.
The shaded map is built on territory; the table is built on the speaker. Shading needs a country code, so it inherits everything a code cannot carry: a historical speaker appears inside its successor's borders, a large state draws the eye for being large, and 0 delegations have no outline at this scale and are marked with a dot. Where two measurable speakers share a code, the country is outlined and left blank rather than given one of their two rates. The circles and this table carry none of these problems.